The role of polysubstance use in the development, maintenance, and treatment of stimulant use disorders

Abstract

Stimulant use disorders represent a significant public health challenge, with no U.S. Food and Drug Administration (FDA) pharmacotherapies currently available. A growing concern is that stimulant use rarely occurs in isolation. Instead, it often involves sequential or simultaneous use of multiple substances. This review explores the mechanistic, epidemiological, clinical, and preclinical dimensions of polysubstance use involving stimulants, particularly cocaine and methamphetamine. Key gaps in the existing literature are identified to underscore the critical need for polysubstance use research across epidemiological, clinical, and preclinical domains. Additionally, the review highlights the importance of fostering interdisciplinary collaborations across these domains to inform the development of more effective interventions for stimulant use disorders and to mitigate the widespread harm caused by substance use globally.

Significance Statement

Stimulant use rarely occurs in isolation and is frequently accompanied by polysubstance use, which increases health risks and complicates prevention and treatment strategies. This review highlights critical gaps in research examining polysubstance use involving stimulants and emphasizes the urgent need for the study of the co-use of drugs and interdisciplinary collaboration. Addressing these gaps is essential to inform the development of effective interventions for stimulant use disorders.

I. Introduction

A. The stimulant use crisis

In the United States and globally, the substance misuse crisis continues to evolve and cause devastating consequences for individuals, families, and communities.^1,2 While opioids have dominated much of the public’s attention due to significant increases in opioid-related overdoses in recent years, stimulants are playing an increasingly dangerous role in the drug epidemic.³ In fact, the rising use of stimulants, often in conjunction with other licit and illicit drugs, has been termed the “silent” epidemic since the marked resurgence of stimulant use and stimulant-associated harm has not received the widespread attention that opioid use has.^4,5 Drug overdose deaths involving stimulants have increased by more than 300% from 2013 to 2019,⁶ and between 2021 and 2024, more than half of all overdose deaths involved stimulants.⁷ Moreover, unlike opioids, there are currently no available medication-assisted treatments for the use of stimulants, which makes the management of stimulant use disorders (StUDs) especially challenging in the clinical setting.^8,9

While stimulant use can result in adverse consequences for mental, neurological, and cardiovascular health, the use of other drugs with stimulants can further exacerbate these risks, potentially leading to increased cardiovascular strain, heightened neurotoxicity, and a greater likelihood of developing mental health disorders such as anxiety, depression, or psychosis.^10,11 Recently, there has been a growing appreciation that although StUDs have been historically conceptualized in isolation as class-specific disorders (eg, cocaine use disorder [CUD]), recent research has suggested that most people who use stimulants engage in polysubstance use (eg, using cocaine and alcohol together).^12,13 Thus, this review discusses the mechanistic, epidemiological, clinical, and preclinical aspects of polysubstance use involving stimulants, primarily cocaine and methamphetamine, highlighting the need to integrate the co-use of other drugs into research and treatment frameworks to advance our understanding of factors mediating continued stimulant use, the long-term consequences, and strategies to improve therapeutic outcomes. Particular emphasis will be placed on animal behavioral pharmacology studies to better elucidate gaps in the preclinical literature on stimulant-related polysubstance use and to explore their implications for study designs and intervention development.

B. Polysubstance use and its impact

In the broadest sense, polysubstance use is defined as the use of more than 1 substance. However, in the scientific community, polysubstance use remains poorly defined, likely due to complexities associated with drug use patterns and motivations for co-using substances. A recent review proposed that to fully define polysubstance use, researchers need to consider the substances involved, the timing of drug co-use, and the intent of use.¹⁴ In fact, polysubstance use should likely focus on capturing short-term behaviors rather than being used to describe multiple substances used across longer periods of time, such as weeks or months. Within a short timeframe, a distinction is also normally made between 2 patterns of polysubstance use: simultaneous and sequential. While sequential polysubstance use is defined as using multiple drugs during a single episode (eg, drinking alcohol and using cocaine during the same party), simultaneous use often encompasses drug use where multiple substances are mixed and self-administered using the same mode (eg, injecting cocaine and heroin in the same syringe).

Another important factor to consider when defining polysubstance use is the motivation for using multiple drugs.¹⁴ The 2 common reasons given by individuals with StUDs for co-using substances are related to positive and negative reinforcement.^15,16 Positive reinforcement occurs when behavior that leads to the presentation of a stimulus increases the probability of that behavior occurring again. In short, a person may use alcohol when using cocaine because the alcohol potentiates the subjective feelings of euphoria associated with the cocaine high. Negative reinforcement, on the other hand, occurs when the behavior that leads to the removal of a stimulus increases the probability of that behavior occurring again. For example, a person may drink alcohol when using cocaine to mitigate feelings of anxiety or reduce withdrawal associated with cocaine use. Although both motivators increase the probability of drug co-use, treatment practices may differ based on intent. In a sense, if polysubstance use is motivated by positive reinforcement, then it may be necessary to treat the use of multiple drugs. However, if negative reinforcement drives polysubstance use, then a treatment for one drug may ameliorate the use of the other drug(s)¹⁴ (Fig. 1). Another important note is that the motivation for co-using drugs may also impact the pattern of drug use. For instance, if a person with CUD uses alcohol to increase the subjective effects of cocaine, they may use alcohol before and during cocaine use. However, if a person with CUD uses alcohol to decrease cocaine-induced anxiety or later withdrawal symptoms, they may use alcohol primarily after cocaine self-administration.

Fig. 1.

Motivations for drug co-use organized by reinforcement type. The diagram illustrates 2 primary mechanisms underlying co-use behavior: positive reinforcement and negative reinforcement. Each pathway may require different treatment approaches.

While defining and studying polysubstance use is difficult, given the heterogeneity of the substances co-used, patterns of co-administration, and the intent behind co-use, the current drug use crisis is largely impacted by the use of multiple substances.¹⁶ In fact, polysubstance use involving stimulants has multiple negative health implications. For example, individuals who co-use cocaine and other substances often have poorer mental and physical health statuses, more severe substance use problems, and are less responsive to interventions aiming to reduce drug use.^11,17 Another important note is that the stimulant use crisis is closely intertwined with the ongoing opioid use crisis, such that 41.3% of overdose deaths between 2021 and 2024 involved the co-use of stimulants and opioids.⁷ Furthermore, among emergency department visits, the rate of visits involving stimulant and opioid co-use increased significantly each year from 2016 to 2019.18, 19, 20 Given the serious negative consequences associated with stimulant-related polysubstance use, research is needed to examine the underlying causes of co-use and to investigate how polysubstance use influences the efficacy of interventions.

II. Mechanistic basis of polysubstance use

Although epidemiological data, which will be discussed in later sections, support the notion that polysubstance use is the norm, not the exception, in certain populations of people who use stimulants, the pharmacological and neural basis of co-use is still not fully characterized. As discussed above, the principles of positive and negative reinforcement can partially explain the motivations behind using multiple drugs simultaneously or in close succession. However, the behavioral mechanisms driving polysubstance use must be understood in the context of both pharmacodynamic (PD) and pharmacokinetic (PK) interactions among substances. This is because these interactions can enhance or attenuate a drug effect and, consequently, shape patterns of co-use. These PD and PK interactions will be explored in greater detail in this section of the review.

A. PD underpinnings of polysubstance use

In the field of pharmacology, PDs refers to the effects drugs have on the body and their mechanisms of action. In the central nervous system, cocaine and methamphetamine, referred to as “stimulants” in this review, act by blocking monoamine reuptake transporters and/or vesicular transporters. Noradrenergic (NET), dopaminergic (DAT), and serotonergic transporters (SERT) and synapses are the sites of action for stimulants. Peripherally, stimulants primarily act on NET, which increases adrenergic transmission and can increase blood pressure and cardiac output.²¹ Centrally, cocaine blocks NET, DAT, and SERT, which inhibits the reuptake of these neurotransmitters into the presynaptic neuron. This inhibition increases the concentration of monoamines at the synapse, thereby enhancing neural activity associated with these neurotransmitters.²² Methamphetamine, like cocaine, disrupts monoamine reuptake by DAT, NET, and SERT. However, methamphetamine also reverses the function of DAT, NET, and SERT, which also induces the release of these monoamines into the synapse.²³

Although cocaine and methamphetamine influence all these neurotransmitter systems, numerous studies have demonstrated that the reinforcing effects of stimulants can be attributed primarily to their actions on dopaminergic neurotransmission in the mesolimbic pathway in the brain. In fact, numerous pharmacological and behavioral studies indicate that increases in dopaminergic neurotransmission in this pathway are both necessary and sufficient to induce stimulant reinforcement.24, 25, 26 While commonly co-used substances with stimulants, including nicotine, alcohol, cannabis, opioids, and benzodiazepines, also increase dopaminergic neurotransmission in the mesolimbic pathway, studies have shown that other neural processes also contribute significantly to the reinforcing effects of these drugs.^24,27,28 Thus, there are numerous complex PD interactions that occur centrally that may mediate polysubstance use involving stimulants.

One of the most widely discussed PD interactions between stimulants and several commonly co-used drugs involves synergistic or additive increases in synaptic dopamine concentrations in the mesolimbic pathway, particularly in the nucleus accumbens. Microdialysis studies examining the co-use of cocaine and alcohol²⁹ and cocaine and heroin³⁰ have shown that the coadministration of alcohol and heroin with cocaine produced synergic elevations of dopamine in the nucleus accumbens compared with cocaine alone. Moreover, simultaneous administration of cocaine and nicotine and methylphenidate and nicotine synergistically elevated dopamine in the nucleus accumbens.³¹ These types of interactions between stimulants and commonly co-used drugs are likely the neural substrates that perpetuate drug co-use through positive reinforcement. In short, since elevated dopamine levels in the mesolimbic pathway and nucleus accumbens are associated with subjective reports of euphoria, substances like alcohol, heroin, and nicotine may be co-used with stimulants to enhance these positive subjective effects.32, 33, 34

The alleviation of undesirable drug effects through polysubstance use can also be explained by PD interactions between stimulants and commonly co-used drugs. Semistructured interviews with people who use cocaine have found that a significant majority of participants take benzodiazepines to counter the negative effects of a cocaine high, particularly anxiety, paranoia, and insomnia.³⁵ Similarly, individuals who use cocaine reported using alcohol to alleviate cocaine-induced feelings of anxiety.^36,37 Benzodiazepines and alcohol both increase γ-aminobutyric acid (GABA) neurotransmission in the brain, although through different mechanisms.38, 39, 40 GABA is the main inhibitory neurotransmitter in the central nervous system and plays an important role in homeostasis during anxiety by counteracting the effects of corticotropin-releasing hormone on the activation of the stress response system (ie, the hypothalamic-pituitary-adrenal axis).⁴¹ Given that studies have shown that numerous stimulants, including cocaine, can induce anxiety by stimulating the hypothalamic-pituitary-adrenal axis, benzodiazepines and alcohol may be co-used to restore homeostasis through enhanced GABAergic neurotransmission.42, 43, 44, 45 Furthermore, studies examining cocaine and heroin co-use have reported that cocaine alleviates withdrawal symptoms associated with opioid dependence.46, 47, 48, 49 While the exact neural substrates underlying these interactions have not been fully characterized, some proposed mechanisms include cocaine attenuating dopamine depletion in the striatum^50,51 and cocaine decreasing noradrenergic hyperactivity during opioid withdrawal.⁴⁶

A limited number of studies using positron emission tomography (PET) imaging have also begun to examine how stimulant co-use alters dopamine receptor availability in vivo. For example, one study in male rhesus monkeys modeled polydrug use by having animals self-administer cocaine and then, several hours later, consume ethanol (2.0 g/kg) or an ethanol-free solution each day.⁵² The study examined how ethanol consumption modified the effects of cocaine self-administration on brain dopamine D2 and D3 receptors using [¹¹C]raclopride PET imaging and quinpirole-induced yawning. In this temporally separated but same-day co-use model, chronic cocaine self-administration decreased D2/D3 receptor availability in striatal subregions only in monkeys that did not drink ethanol, whereas ethanol-drinking monkeys showed preserved D2/D3 receptor availability but significant increases in D3 receptor sensitivity, as indexed by quinpirole-elicited yawning. These findings suggest that ethanol co-use may modify cocaine’s long-term effects on dopamine receptors and implicates D3 receptors as a potential medication target in individuals who co-use alcohol and cocaine. However, PET studies that directly interrogate polydrug use involving stimulants remain scarce overall, underscoring the need for additional work in this area to clarify how co-use patterns shape brain receptor function. For instance, seminal PET studies, such as the study by Volkow et al⁵³ examining striatal D2/D3 receptor availability in individuals with cocaine dependence, largely excluded participants with substantial polydrug use (allowing only nicotine and caffeine use and moderate alcohol consumption), and no analyses were conducted to evaluate how these commonly co-used substances might have influenced PET outcomes.

Because the primary emphasis of this review is animal behavioral pharmacology studies, this is not an exhaustive description of all the studies examining this topic. However, the above examples highlight that the PD interactions between stimulants and commonly co-used substances underlie polysubstance use involving stimulants. While stimulants primarily exert their reinforcing effects through increased dopaminergic transmission in the mesolimbic pathway, co-use with substances such as alcohol, nicotine, benzodiazepines, and opioids can enhance or mitigate certain drug effects through synergistic or compensatory mechanisms. The amplification of dopamine release in the nucleus accumbens may contribute to the reinforcing effects of drug combinations, whereas modulation of inhibitory neurotransmission, such as enhanced GABAergic activity, could counteract stimulant-induced anxiety. Additionally, the interplay between stimulants and opioids may influence withdrawal symptoms, further driving co-use behaviors. To fully characterize adaptations following polysubstance use, additional in vivo measures of receptor function and availability are needed to directly link PD interactions to brain receptor changes over time. In sum, a deeper understanding of these PD interactions is crucial for the development of effective interventions for individuals engaging in polysubstance use.

B. Pharmacokinetic underpinnings of polysubstance use

In addition to PD interactions between stimulants and commonly co-used drugs, there are also PK interactions that may occur. These interactions occur when one drug alters the absorption, distribution, metabolism, or excretion of another drug and can result in either an increase or a decrease in plasma drug concentrations.⁵⁴ It is important to recognize that many apparent discrepancies in PK findings across studies may be attributable to differences in the route of administration. Oral administration is subject to extensive first-pass hepatic metabolism, which can reduce systemic bioavailability, whereas intravenous and smoked routes bypass this effect and yield faster onset and higher peak plasma concentrations.⁵⁵ These route-dependent factors can therefore alter absorption rates, plasma levels, and metabolite profiles, and should be considered when comparing across studies. For example, while smoked and intravenous administration of cocaine have been shown to result in approximately equivalent plasma concentrations of cocaine, the intranasal and oral routes result in lower plasma concentrations and a slower onset of pharmacological effects.^56,57 However, the addition of other commonly co-used drugs can modify these PK profiles.

1. Co-use with nicotine

As will be described later in this review, co-use of nicotine can potentiate many behavioral effects of cocaine, so examining how nicotine exposure alters PK parameters of stimulants is highly relevant. With that said, there is not much literature on nicotine effects on stimulant PK. One of the challenges involves the route of administration. Quinn et al⁵⁸ noted that the smoking and inhalation routes of administration resulted in the most rapid delivery of nicotine to the brain, while intravenous injection resulted in the maximum amount of bioavailability of an administered drug dose. These differences in PK based on route are important when discussing nicotine research, since humans typically inhale nicotine, while preclinical studies typically use systemic (primarily intravenous or intramuscular) routes.

To better understand the PK parameters of nicotine, Moerke et al⁵⁹ trained rhesus monkeys to discriminate nicotine (0.032 mg/kg i.v.) from saline while collecting serial blood samples to establish a time course of plasma nicotine levels. Although the primary objective was behavior, the study provided valuable PK data: plasma nicotine concentrations peaked within 9 minutes and had a mean half-life of ∼116 minutes, whereas discriminative stimulus effects dissipated within 30 minutes. This dissociation highlights how plasma nicotine concentrations can persist beyond its acute behavioral effects. These findings, while derived from drug discrimination studies, inform our understanding of the absorption and clearance dynamics of nicotine, which are integral to assessing its PK interplay with stimulants.

Vieira-Brock et al⁶⁰ noted that in preclinical studies, repeated methamphetamine administration caused persistent deficits in the dopaminergic system resembling Parkinson’s disease. The investigators noted that many people who use methamphetamine smoke cigarettes and were interested in whether nicotine affected methamphetamine-induced dopaminergic deficits. Relevant to the study was the timing of nicotine exposure. Because smoking frequently begins at a young age in people, in their model, rats drank nicotine (available in the drinking water) from adolescence (postnatal day [PND] 40) through adulthood (PND 96), with methamphetamine administered at a dosing regimen shown to disrupt dopaminergic function at PND 89. Nicotine’s neuroprotective effects were age-dependent. While the major dependent variables in this study⁶⁰ were changes in brain nicotine and dopamine receptor densities, of relevance to this section, the investigators also examined whether nicotine exposure altered methamphetamine PK by assessing concentrations of methamphetamine and its metabolite, amphetamine, in rats exposed to tap or nicotine water. They found no differences in methamphetamine or amphetamine concentrations in rats pre-exposed to nicotine water or tap water, suggesting that the neuroprotective effects of nicotine were not due to changes in PK.

2. Co-use with alcohol

Similar findings to those reported for cannabis have been reported in the literature examining stimulant and alcohol co-use. A study of humans examining the interactions between oral ethanol and intranasal cocaine in short succession found that cocaine plasma levels were higher when alcohol was administered compared with cocaine alone, proposing there was an inhibition of the hepatic metabolism of cocaine in the presence of alcohol.⁶¹ The study also found that absorption of cocaine was faster when co-used with alcohol and suggested that alcohol may have resulted in a vasodilation of the nasal mucosa and thus increased cocaine intranasal absorption.⁶¹ In rodents, intraperitoneal dosing of cocaine in conjunction with an intravenous infusion of ethanol resulted in a faster rate of cocaine absorption and, in turn, higher plasma cocaine concentrations and higher cocaine brain levels.⁶² Moreover, in rats, intraperitoneal ethanol was shown to increase the elimination half-life of intraperitoneal cocaine by inhibiting cocaine metabolism and increasing cocaine tissue distribution.⁶³ Together, these findings suggest that ethanol could enhance the absorption, distribution, and elimination of cocaine under a variety of experimental conditions.

Interestingly, a series of human studies suggests that the PK interactions between cocaine and alcohol vary based on the order of administration. In one study,⁶⁴ participants consumed ethanol or a placebo over 30 minutes, and 15 minutes after the end of drinking, received intranasal cocaine HCl or a placebo. Cocaine administration did not alter blood ethanol concentrations or ratings of ethanol intoxication, but ethanol produced significant increases in cocaine plasma concentrations, ratings of cocaine “high,” and heart rate. In a companion study⁶⁵ using the reverse order of administration, intranasal cocaine was given first, and ethanol or a placebo was consumed 30 minutes later. Under these conditions, ethanol given after cocaine did not increase cocaine plasma concentrations, augment the subjective or cardiovascular effects of cocaine, or change blood ethanol levels or ratings of ethanol intoxication. These findings suggest that ethanol may modify the PK of cocaine under specific dosing conditions; however, similar effects have not been observed with other stimulants. For example, studies examining intravenous methamphetamine and oral ethanol interactions in humans found that ethanol did not modify methamphetamine PK.⁶⁶ Thus, there appear to be substance-specific effects of alcohol on stimulant PK.

Another important note about cocaine and alcohol co-use is that this combination results in the formation of a metabolite in the liver called cocaethylene. Cocaethylene has a similar pharmacological profile to cocaine at the monoamine transporter sites in the human brain.⁶⁷ Studies have also supported the premise that this metabolite is reinforcing, likely due to its high affinity for DAT.68, 69, 70, 71 In addition, studies have shown that cocaethylene is more toxic than cocaine. In fact, cocaethylene carries an 18–25-fold increase in risk of death due to seizures, liver damage, or heart failure over cocaine alone.^72,73 Moreover, cocaethylene is eliminated more slowly than cocaine, which may prolong subjective reports of euphoria.⁷⁴ Importantly, the amount of cocaethylene formed depends on the route of cocaine administration following oral ethanol administration. In one human study,⁷⁵ approximately one-third of oral cocaine was converted to cocaethylene compared with about one-quarter following intravenous administration and less than one-fifth after smoked cocaine, consistent with a larger contribution of first-pass metabolism for the oral route. Taken together, these data highlight the complexity of cocaine-alcohol polydrug use and underscore the importance of carefully considering PK interactions when interpreting both experimental and clinical outcomes.

3. Co-use with cannabis

Studies have shown that people who smoke cannabis in close succession to intranasal cocaine had increased plasma cocaine levels and bioavailability relative to those who did not smoke cannabis prior to cocaine.^76,77 One proposed mechanism was that cannabis-induced vasodilation in the nasal mucosa allowed for increased absorption of intranasal cocaine.^76,77 However, another study examining the effects of smoked cannabis on smoked cocaine found that cannabis doses that were higher than those examined in the previous study decreased plasma levels of cocaine and its metabolites.^77,78 This may suggest that the PK interactions between cannabis and cocaine vary based on the route of cocaine administration and the cocaine dose.

In preclinical models, intraperitoneal pretreatments with cannabidiol (CBD) and tetrahydrocannabinol (THC), 2 major constituents of cannabis, 1 hour prior to an intraperitoneal injection of cocaine, significantly increased cocaine levels in the brain. One hypothesis proposed was that THC and CBD interact with plasma-binding proteins, allowing more drug to pass through the blood-brain barrier.⁷⁹ Supporting evidence for this hypothesis has been mixed, with some studies showing that THC induces blood-brain barrier dysfunction^80,81 and others showing that THC has protective effects on blood-brain barrier integrity.⁸² While the exact mechanisms underlying these interactions have not been elucidated, the data do support the premise that cannabis, under certain conditions, can influence the absorption and bioavailability of cocaine. Investigations of the PK interactions between cannabis and other stimulants, including methamphetamine, have been more limited. While studies have shown that cannabis co-use with methamphetamine could potentiate neurocognitive impairments, no studies, to our knowledge, have directly investigated how cannabis influences the absorption, distribution, metabolism, or excretion of methamphetamine.⁸³

4. Co-use with opioids

This section focuses primarily on cocaine. In relation to cocaine + opioid polysubstance use, there are 2 general types of questions being asked: how does cocaine affect opioid PK, and conversely, how does opioid use affect cocaine PK? Cocaine use frequently occurs in people with opioid use disorders (OUDs), including those in treatment. For example, one study examined 495 people who used heroin and were assigned to a treatment study in New South Wales, Australia, and noted that nearly all of them (∼91%) had a history of cocaine use.⁸⁴ The investigators noted that any cocaine use during the study was associated with poorer treatment outcomes. Research by McCance-Katz et al^85,86 has shown that chronic cocaine exposure significantly decreased buprenorphine and methadone plasma concentrations. One possible explanation provided involves cocaine and one of its metabolites, benzoylecgonine (BE), both of which are vasoconstrictors, which could affect opioid PK. Of note, neither study described the route or frequency of cocaine use, which implied that it was the chronic nature of exposure rather than the acute effects that influenced opioid PK. The investigators concluded that people who use cocaine regularly may negatively impact treatment outcomes for OUD. In an extension of these findings, Tetraul et al⁸⁷ examined plasma methadone and buprenorphine concentrations in individuals who were currently using cocaine versus those who were cocaine abstinent, as a function of human immunodeficiency virus (HIV) status. They found that peak buprenorphine concentrations were not different between people who used cocaine and cocaine-abstinent individuals, while peak methadone (both R- and S-isomers) concentrations were lower in individuals who used cocaine. Interestingly, they found that methadone concentrations were not decreased in people who used cocaine and were HIV-positive compared with cocaine-abstinent individuals, suggesting that either HIV or its treatment attenuated the effects of cocaine on methadone plasma levels.

In a recent preclinical study, Wei et al⁸⁸ examined metabolism following intraperitoneal administration of cocaine, heroin, and cocaine and heroin (ie, speedball) in rats. In this study, rats were injected with either 20 mg/kg cocaine (N = 9), 10 mg/kg heroin (N = 8), or a combination of 20 mg/kg cocaine and 10 mg/kg heroin (N = 12). Three metabolites of cocaine were examined: ecgonine methyl ester and BE, both metabolized in plasma, and the liver metabolite norcocaine. Blood concentrations of cocaine, ecgonine methyl ester, BE, and norcocaine were significantly higher when heroin was coinjected with cocaine. For heroin, the results differed slightly. In plasma, heroin is converted to 6-monoacetylmorphine (6-MAM), which is then converted to morphine; the morphine is further conjugated to morphine-3-glucuronide and morphine-6-glucuronide. Of significance was the observation that blood concentrations of 6-MAM were higher in the heroin-only group compared with the speedball group, but blood morphine concentrations were higher in the speedball group, suggesting that metabolism of 6-MAM was accelerated when cocaine was co-administered. It remains an empirical question as to how these effects would differ if heroin and/or cocaine were smoked rather than administered systemically (for example, see Cook⁸⁹). One important note is that while all rats in the cocaine group showed hyperactivity and none showed toxicity, all rats treated with heroin showed signs of toxicity, including sedation and respiratory depression, described as a “nonfatal overdose.” However, when cocaine and heroin were combined, 4 of the 12 rats died (33%).

There are very few studies examining PK interactions between opioids and methamphetamine.^90,91 However, because methamphetamine and various opioids, such as oxycodone and codeine, use the liver enzyme CYP2D6 as the main route of metabolism, it is possible that co-using these stimulants with opioids can slow metabolism and perhaps influence the duration and intensity of certain drug effects.91, 92, 93 More studies investigating these interactions are warranted to further elucidate the potential mechanisms underlying co-use.

C. Summary and conclusions

While not every PK interaction between substances and stimulants is cited here, the discussed studies support the notion that these interactions play a crucial role in shaping the effects of stimulant co-use. By altering absorption, distribution, metabolism, and excretion, these interactions can lead to increased plasma drug concentrations, prolonged drug effects, and enhanced toxicity. Mechanistically, the co-use of substances like cannabis and alcohol with stimulants can facilitate increased absorption via vasodilation, altered distribution through interactions with the blood-brain barrier, and slowed metabolism, leading to higher plasma and brain drug levels. These interactions may enhance the subjective effects of stimulants and contribute to prolonged stimulant availability in the system. Collectively, differences in route of administration, including the presence or absence of first-pass metabolism, likely account for some of the variability in reported stimulant PK interactions, underscoring the need for standardized comparative designs. Importantly, these PK interactions also intersect with the positive and negative reinforcement mechanisms described earlier. Increases in drug bioavailability or duration of action can amplify positive reinforcement, encouraging co-use to enhance euphoria or prolong intoxication. Alternatively, when one substance alters metabolism or clearance in a way that diminishes withdrawal or other unpleasant effects of another drug, negative reinforcement may sustain co-use by providing relief from those aversive states. Thus, PK processes may underlie both the pursuit of enhanced reward and the avoidance of discomfort, which together reinforce polysubstance use patterns.

III. Epidemiological findings of stimulant polysubstance use

As previously noted, there has been increasing recognition that although StUDs have traditionally been conceptualized as distinct class-specific conditions, emerging epidemiological research indicates that most individuals who use stimulants also engage in polysubstance use.^12,13 Thus, this section reviews key epidemiological findings that characterize the patterns, prevalence, and factors associated with sequential and simultaneous use of numerous other drugs with cocaine and methamphetamine. Emphasis is placed on understanding how these patterns vary across populations as well as their associations with an increased risk of adverse health outcomes, including psychiatric comorbidities, infectious disease, and overdose. By synthesizing current epidemiological data, this section aims to contextualize the scope and significance of stimulant-related polysubstance use within broader substance use trends.

A. Polysubstance use involving cocaine

Several studies have examined the prevalence of polysubstance use in individuals who use cocaine. One study found that 76% of people who use cocaine reported binge drinking, 85% of people who use cocaine smoked cannabis, and 80% of people who use cocaine used nicotine-containing products.²⁸ Moreover, individuals who use cocaine also reported co-using prescription stimulants (28%), prescription opioids (25%), prescription sedatives (24%), and heroin (8%).²⁸ Additionally, another study, which collected data from almost 70,000 admissions to publicly funded treatment programs, reported that 87.8% of people who use cocaine reported polysubstance use. More in-depth analyses demonstrated that the most common 2-substance combination was alcohol + cocaine, followed by alcohol + cannabis, cocaine + cannabis, and alcohol + opioids. Moreover, among individuals who reported using 3 substances, the most common combination was alcohol + cocaine + cannabis, followed by alcohol + cocaine + opioids.²⁷ Another study examined hour-level patterns of polysubstance use involving cocaine, alcohol, and cannabis and found that the most common pattern of co-use involved simultaneous cocaine and alcohol use in the late afternoon and evening. Another commonly reported pattern involved extensive cannabis use and then co-use of cocaine, alcohol, and cannabis.⁹⁴ It is important to note, however, that significant heterogeneity is present in the reported prevalence of polysubstance use in people who use cocaine in epidemiological studies assessing commonly co-used substances (alcohol and cannabis), with ranges between 24% and 98% for alcohol co-use and 12% and 76% for cannabis use.⁷⁷

Furthermore, demographic predictors of polysubstance use involving cocaine have been mixed, likely due to variability in the target population, study location, year, and demographic makeup of each study.⁷⁷ For instance, while in the above study,²⁷ there were no biological sex differences noted in the probability of engaging in polysubstance use, other studies have found that men who use cocaine were more likely to engage in polysubstance use.^95,96 While biological divergence may account for differences in the prevalence of polysubstance use involving cocaine, it is also possible that sociocultural factors underlie these differences. Some studies have reported that African Americans who use cocaine are more likely to engage in polysubstance use compared with other ethnicities.^27,97 Interestingly, one of these studies further suggests that this relationship appears to be independent of income or educational level and that differences may be attributed to cultural and racial-ethnic factors rather than to disparities in economic status.⁹⁷ However, another study evaluating the prevalence of polysubstance use involving cocaine among various ethnic groups found no differences based on ethnicity.⁹⁵ Another finding reported that sexual orientation influenced the probability of polysubstance use involving cocaine, and that identifying as nonheterosexual increased the odds of co-using cocaine with other substances, such as cannabis.^98,99

Mental and physical health status have also been identified as being associated with polysubstance use. For instance, one study found that compared with individuals with CUD alone, those who had concurrent CUD and alcohol use had higher depression scores, more severe CUD symptoms, and were more likely to experience a paranoid psychosis.¹⁰⁰ Another study found that individuals who engaged in polysubstance use, most commonly cannabis and cocaine, had a higher likelihood of reporting lifetime major depressive disorder and generalized anxiety disorder than those who had no illegal drug use in the past year.¹⁰¹ In addition, alcohol-dependent individuals with a history of sexually transmitted infections were more likely to co-use cocaine than those with no history of sexually transmitted infections.¹⁰¹ Other physical health issues, such as untreated periodontitis and liver fibrosis, have been found to be higher among individuals who co-use substances with cocaine compared with those who only use cocaine.^102,103

Finally, the co-use of substances with cocaine may increase the risk of mortality. In fact, after controlling for overdoses caused by a single drug, one study found that more than half of accidental overdose deaths over an 8-year period in New York City could be attributed to cocaine, alcohol, or opioids used in combination.¹⁰⁴ Another study assessed the mortality risk of patients in the emergency department and found that combined cocaine and cannabis use resulted in higher death rates compared with other patients, likely due to an increased risk of cardiac events among individuals who used cocaine and also smoked cannabis.¹⁰⁵ Cocaine-involved deaths are also rising due to the co-administration of potent synthetic opioids like fentanyl.^106,107 However, it remains unclear whether this increase was primarily driven by an intentional co-administration of cocaine and synthetic opioids or whether these synthetic opioids were mixed into cocaine products without the knowledge of the person using cocaine. Regardless of intent, the presence of multiple substances, especially highly potent ones like fentanyl, amplifies the risk of fatal outcomes when using cocaine.¹⁰⁷

B. Polysubstance use involving methamphetamine

Polysubstance use involving methamphetamine is also ubiquitous. One study found that over half of outpatients treated for methamphetamine use disorder (MUD) concurrently used alcohol (63%) and cannabis (59%).¹⁰⁸ In addition, 10% of people who used methamphetamine co-used prescription opioids, 9% co-used tranquilizers, and 6% co-used heroin.¹⁰⁸ Nicotine co-use is also highly prevalent.¹⁰⁹ However, more recent studies have found a much higher prevalence of methamphetamine and opioid co-use, with one study reporting co-occurring OUD in over 90% of patients with MUD.¹¹⁰ This suggests that the prevalence of methamphetamine and opioid co-use is growing. Similar to studies examining polysubstance use involving cocaine, those examining methamphetamine and alcohol polysubstance use have found a positive association between methamphetamine use and the risk of binge drinking alcohol.¹¹¹ To our knowledge, no studies have directly evaluated the temporal patterns of polysubstance use involving methamphetamine.

However, numerous studies have reported demographic and health variables related to methamphetamine polysubstance use. One study found that among people who inject drugs, those who were younger, recently incarcerated, had an opioid overdose, injected drug/drug combinations more than once per day, and shared syringes were more likely to co-use methamphetamine and opioids compared with using the drugs in isolation.¹¹² Moreover, individuals who used methamphetamine who identified as nonheterosexual, had less than a high school education, were unemployed, uninsured, and homeless were more likely to fall into the co-use category.¹¹² Another study found that in a population of people who used methamphetamine, less than 25% of participants used a single drug, and the highest percentage of individuals co-used methamphetamine and cannabis (39%). In this population, people who engaged in co-use were more likely to be male, unmarried, unemployed, and White.⁸⁵ In addition, greater depressive symptomatology was reported in people who engaged in polysubstance use of methamphetamine and cannabis compared with those who used methamphetamine alone and drug-abstinent individuals.¹¹³ Moreover, reports suggest that individuals who co-use methamphetamine and opioids are more likely to report having hepatitis C than those who use methamphetamines alone, and more severe mental illnesses than those who use opioids alone.^112,114

Like the patterns seen with cocaine, the co-use of other substances alongside methamphetamine significantly increases the risk of mortality. In the United States, methamphetamine-related deaths have risen sharply over the past 2 decades.^115,116 A growing body of research suggests that this trend is closely linked to the increased co-use of fentanyl. Synthetic opioids, like fentanyl, are also increasingly found as contaminants in the unregulated drug supply, which may be a major driver of methamphetamine-related overdose deaths.^115,117,118 Research on the impact of other commonly co-used drugs, such as cannabis and alcohol, on methamphetamine-related mortality has been limited. However, one study found that patients who co-used cannabis and methamphetamine had a 20-fold increase in the probability of a diagnosis of methamphetamine-associated heart failure.¹¹⁹ To our knowledge, however, no studies have shown whether the co-use of alcohol potentiates the risk of methamphetamine-related mortality. These gaps in the literature highlight a critical need for further research on how the co-use of substances like cannabis and alcohol may contribute to the rising burden of methamphetamine-related deaths.

C. Summary and conclusions

The above-mentioned epidemiological findings (summarized in Table 1) underscore the pervasive and complex nature of polysubstance use involving stimulants, such as cocaine and methamphetamine. Across substances, patterns of co-use – often involving alcohol, cannabis, opioids, and prescription medications – are the norm rather than the exception, with evidence pointing to specific combinations, temporal patterns, and demographic associations. Notably, stimulant-related polysubstance use is associated with elevated risks for a range of adverse outcomes, including psychiatric comorbidities, infectious disease, cardiovascular complications, and overdose. Demographic shifts, such as the rising prevalence of stimulant-opioid co-use among White individuals and elevated co-use rates among nonheterosexual and socioeconomically disadvantaged populations, suggest that polysubstance use is shaped by broader sociocultural and structural forces.

Table 1.

Summary of key epidemiological findings on polysubstance use among individuals with CUD and MUD

Domain	Cocaine-Related Findings	Methamphetamine-Related Findings
Prevalence of polysubstance use	Highly prevalent, with one study noting that almost 90% of people with CUD report polysubstance use.27	Highly prevalent, with one study noting that over 50% of people with MUD report polysubstance use.108 This is likely growing, especially for the co-use with opioids.109
Most common co-used substances	Alcohol, nicotine, cannabis, opioids, and prescription sedatives.27,28,77,94	Alcohol, nicotine, cannabis, opioids, and prescription tranquilizers.108, 109, 110, 111
Temporal patterns of co-use	Highly heterogeneous, often involving simultaneous and sequential co-use across numerous hours.94	No studies evaluating temporal patterns.
Demographic correlates	Some mixed data suggest higher prevalence in African Americans, men, and nonheterosexual individuals.27,95, 96, 97, 98, 99	Higher prevalence is related to numerous demographic variables, including being young, White, unmarried, unemployed, homeless, and nonheterosexual.85,112
Psychiatric comorbidities	Higher likelihood of depression, more severe CUD, more likely to experience psychosis, and have an anxiety disorder.100,101	Higher likelihood of depression and several mental illnesses.112,113
Infectious disease and other medical outcomes	Higher risk of STIs, periodontitis, and liver fibrosis.101, 102, 103	Higher risk of hepatitis C.112,114
Mortality and overdose	Higher risk of fatal outcomes.104, 105, 106, 107	Higher risk of fatal outcomes.115, 116, 117, 118, 119

STI, sexually transmitted infection.

However, it is critical to note that most of these findings were derived from cross-sectional studies, limiting the ability to draw causal inferences. It remains unclear whether psychiatric or medical issues increase the likelihood of engaging in polysubstance use, or whether sustained co-use drives the emergence of these conditions. These complexities highlight the urgent need for longitudinal and mechanistic studies to disentangle causality and guide more effective interventions. Ultimately, stimulant polysubstance use reflects a dynamic and multidimensional public health challenge that requires coordinated prevention, harm reduction, and treatment strategies informed by the interplay of substance use behaviors, social determinants, and health outcomes.

IV. Findings in humans of stimulant polysubstance use

While epidemiological studies highlight the widespread prevalence and factors that impact stimulant polysubstance use, clinical research provides important insights into how these combinations affect subjective reports of drug effects and other behavioral outcomes. Clinical research on stimulant polysubstance use can be broadly divided into 2 categories: human laboratory studies and clinical trials. Human laboratory studies involve controlled drug administration and provide critical mechanistic insights into how combinations of stimulants with other substances alter subjective, behavioral, and physiological effects. In contrast, clinical trials focus on treatment outcomes in naturalistic settings. Although the 2 study categories differ in scope, both provide valuable information: human laboratory data help explain the pharmacological interactions underlying co-use, whereas clinical trials highlight how co-use patterns may influence treatment responses. To clarify these distinctions, the following subsections are organized by co-use substances, with human laboratory studies and clinical trials discussed separately, if possible. However, because research designs vary across substances, not all subsections contain both types of data; in some cases, only 1 form of evidence is available.

This section synthesizes key clinical findings on the effects and implications of polysubstance use involving cocaine and methamphetamine, with a focus on common co-use substances such as nicotine, alcohol, cannabis, and opioids. Importantly, this section is not an exhaustive survey of all potential pharmacotherapies for CUD or MUD. Instead, we deliberately focus on studies that explicitly address polysubstance use, specifically those evaluating how co-use of another substance modulates key outcomes. Trials without polysubstance analyses are not covered to preserve a tight focus on the clinical relevance of polysubstance use.

A. Stimulants and nicotine

1. Cocaine

a. Human laboratory findings

Several studies have investigated the interactions between cocaine and nicotine use. In one study, individuals with a history of smoking both cocaine and cigarettes were exposed to cocaine-related cues by watching a 5-minute video and then handling cocaine paraphernalia for 5–7 minutes. Participants who received a nicotine transdermal patch prior to this cue exposure reported greater cocaine craving than those who received a placebo patch.¹²⁰ Similarly, another study using a range of questionnaires in people who actively use cocaine and smoke cigarettes found that cocaine use strongly increased the urge to smoke and intensified cigarette cravings. In addition, cigarette smoking modestly heightened subjective ratings of feeling high and the desire for more cocaine.¹²¹

A related study in healthy volunteers examined how acute intranasal cocaine affected cigarette smoking over a 3-hour period. Results showed that after cocaine administration, participants smoked their first cigarette sooner, had shorter intervals between cigarettes, and smoked more cigarettes overall compared with participants who received placebo cocaine. In another experiment, urine tests revealed that cotinine levels, a marker of nicotine use, were higher on days when participants tested positive for cocaine, suggesting that cocaine use was associated with increased smoking.¹²² This finding is consistent with other research showing that smoking frequency is higher during episodes of cocaine use and that both smoking and tobacco cravings decrease during periods of verified cocaine abstinence.¹²³

Pharmacodynamic interactions between cocaine and nicotine could explain the above findings. Because both cocaine and nicotine increase dopaminergic neurotransmission, co-use may amplify reinforcement and craving through additive activation of mesolimbic dopamine pathways.³¹ Thus, this combination may enhance self-administration and craving and perpetuate co-use through positive reinforcement mechanisms.

b. Clinical study findings

Additional studies have examined whether reducing the use of one substance affects the use of the other. For example, one study found that although cocaine use decreased significantly during a 12-week outpatient treatment program, the number of cigarettes smoked per day remained unchanged during treatment and follow-up.¹²⁴ In contrast, another study examining the reverse relationship found that participants with CUD who achieved smoking abstinence during a 10-week trial were more likely to remain abstinent from cocaine than those who continued smoking. Interestingly, this effect was not observed in participants dependent on methamphetamine.¹²⁵ Together, these findings suggest that while reducing cocaine use alone may not impact cigarette smoking, helping individuals with CUD to quit smoking might improve cocaine abstinence outcomes. This consideration may be especially important for clinicians who want to treat CUD in patients who are actively using nicotine-containing products.

While few clinical trials have directly tested pharmacotherapies aimed at reducing both cocaine and nicotine co-use, most clinical trials of CUD do not exclude participants who smoke or use nicotine-containing products. However, nicotine use is rarely included as a covariate in statistical analyses, making it difficult to draw clear conclusions about whether an intervention’s effectiveness differs by smoking/nicotine use status.126, 127, 128 This is an important limitation. If nicotine use reduces the efficacy of potential treatments but is not systematically examined, there is a risk of overlooking crucial information that could support more personalized and effective treatment approaches for CUD.

2. Methamphetamine

a. Clinical study findings

Although no studies have directly examined how nicotine affects the subjective experience of methamphetamine use, some research has explored how cigarette smoking may influence the effectiveness of pharmacotherapies for MUD. One study found that smoking status at the start of the clinical trial did not affect participants’ responsiveness to bupropion treatment compared with placebo.¹²⁹ However, among those in the placebo group, heavier cigarette smoking was linked to a higher likelihood of methamphetamine use during the trial.¹²⁹ Since this relationship was not found in the group who received bupropion, it suggests that bupropion may help disrupt the link between cigarette smoking and continued methamphetamine use. Another study found that a behavioral intervention targeting methamphetamine use (ie, contingency management) had a positive impact on tobacco smoking. Participants who received contingency management for MUD were 140% more likely to submit a nicotine-negative sample, indicating that addressing methamphetamine use through behavioral strategies may also promote smoking cessation.¹³⁰

In contrast, several studies suggest that reductions in cigarette smoking may not be directly related to reductions in methamphetamine use. For example, one study showed that while bupropion did not reduce methamphetamine-positive urine drug screens compared with controls, it did help participants reduce cigarette smoking. In this context, the results suggest that reductions in cigarette smoking did not correspond to reductions in methamphetamine use.¹³¹ Moreover, the route of methamphetamine administration (ie, smoked, nasal, injection, or oral) did not influence the outcomes.¹³¹ A similar pattern was observed with varenicline: while it was effective at reducing cigarette smoking among people with MUD who smoked, it did not reduce methamphetamine use.¹³² In that study,¹³² the route of methamphetamine administration was not evaluated. Together, these findings suggest that under certain conditions, cigarette smoking and methamphetamine use may be distinct behaviors and that reducing one does not necessarily affect the other. Moreover, the limited cross-efficacy of these medications likely reflects their pharmacodynamic selectivity. Both bupropion and varenicline modulate nicotinic acetylcholine receptor signaling, which may target mechanisms central to nicotine use but largely peripheral to methamphetamine use.^133,134

3. Summary

In summary, converging evidence demonstrates important interactions between cocaine and nicotine. For example, nicotine can heighten cocaine craving in response to cues, and cocaine increases the urge to smoke and the amount smoked. While reducing cocaine use does not appear to affect smoking behavior, smoking cessation may improve treatment outcomes for CUD. This is an important insight for treatment planning. By contrast, for methamphetamine, the available data focus more on treatment than on nicotine’s acute subjective effects. In general, the data suggest that reducing smoking in patients who co-use nicotine and methamphetamine does not consistently reduce methamphetamine use. Thus, nicotine status appears clinically salient for CUD trials, whereas the link between nicotine use and MUD is less defined.

B. Stimulants and alcohol

1. Cocaine

a. Human laboratory findings

A series of early controlled studies explored how alcohol alters the effects of cocaine and vice versa, providing important insights into why these substances are frequently co-used. In a double-blind controlled trial, researchers examined the impact of administering 100 mg of intranasal cocaine to non-dependent healthy volunteers who were acutely intoxicated with alcohol (1.0 g/kg).⁶¹ During each session, participants consumed an alcoholic beverage over 30 minutes, followed by snorting cocaine. The results showed that the combination of alcohol and cocaine decreased subjective feelings of drunkenness while increasing cocaine-induced euphoria. Interestingly, cocaine also appeared to counteract alcohol-related declines in psychomotor performance. Notably, the study identified a key PK interaction: when alcohol was consumed first, cocaine plasma levels were significantly higher.⁶¹ A similar experiment found comparable results.¹³⁵ Alcohol consumed over 30 minutes followed by 100 mg of intranasal cocaine increased pleasurable subjective effects compared with cocaine alone. The combination also reduced alcohol-induced sedation, though feelings of drunkenness remained unchanged. Importantly, co-use again raised cocaine concentrations in the blood and led to the synthesis of cocaethylene.¹³⁵

These findings have been replicated across multiple studies.136, 137, 138, 139, 140 A comprehensive review¹³⁹ concluded that ethanol consistently enhances the euphoric effects of cocaine in humans, while cocaine consistently offsets alcohol-induced impairments in learning, psychomotor skills, and driving performance. The combination also has additive effects on heart rate and blood pressure, which may contribute to an increased risk of cardiotoxicity.¹³⁹ Studies also support the notion that when alcohol is co-used with cocaine, the euphoric effects are prolonged, which may be attributed to the formation of cocaethylene.¹⁴¹ To date, only 1 study has evaluated whether there are biological sex differences in these effects and found that responses to cocaine, alcohol, and their combination did not differ significantly between men and women in terms of PK measures or subjective effects.¹⁴²

b. Clinical study findings

Several clinical studies have also evaluated the efficacy of various interventions when cocaine and alcohol are co-used. One study examined the efficacy of the DAT blocker modafinil for the treatment of CUD in a double-blind, placebo-controlled trial. They found that although there was no significant difference between the modafinil and placebo groups in terms of cocaine non-use days over a 12-week period, modafinil was efficacious in a subgroup of patients who did not have a history of alcohol dependence.¹²⁷ Similarly, another trial that excluded individuals with a history of alcohol use found that participants treated with modafinil were more likely to maintain at least 3 weeks of abstinence from cocaine and reported lower craving levels than those on placebo. These findings suggest that the co-use of alcohol may diminish the effectiveness of some treatments that could otherwise benefit individuals with CUD.¹⁴³ As described earlier, this reduced efficacy could reflect chronic neuroadaptations to dopamine receptor function that are exacerbated when cocaine and alcohol are co-used or acute PD interactions in which addictive or synergistic increases in dopamine release during co-use blunt the therapeutic impact of dopaminergic or agonist-based treatments.

Additionally, several studies have examined disulfiram, an FDA-approved medication for alcohol use disorder (AUD), as a treatment for patients with co-occurring CUD and AUD. These studies have generally found positive results showing that disulfiram can help reduce both alcohol and cocaine consumption.144, 145, 146, 147, 148 One study found that disulfiram was most effective at reducing cocaine use among patients who were not alcohol-dependent at baseline and who maintained complete abstinence from alcohol during treatment. This suggests that while disulfiram can be beneficial for individuals who co-use cocaine and alcohol, its effectiveness may be greater in those who use cocaine alone.¹⁴⁹ Disulfiram acts primarily by inhibiting aldehyde dehydrogenase, leading to aversive reactions to alcohol consumption due to acetaldehyde accumulation.¹⁵⁰ However, it also inhibits dopamine β-hydroxylase, which converts dopamine to norepinephrine, thereby increasing central dopamine and decreasing norepinephrine.^151,152 This dopaminergic modulation may reduce cocaine use by restoring normal dopaminergic functioning in hypodopaminergic individuals with CUD.¹⁵² Similar to the findings of modafinil, the reduced efficacy observed in individuals who co-use alcohol may reflect chronic neuroadaptations or acute pharmacodynamic interactions, in which alcohol limits the effectiveness of dopaminergic or agonist-based therapeutics for CUD. It is also important to note that, because there are significant concerns about patients taking their medication, its success depends on patient compliance and, ideally, supervised administration.¹⁵³

Clinical studies have also examined the efficacy of naltrexone, an FDA-approved medication for the treatment of AUD, for cocaine and alcohol co-use, and results have been equivocal. For instance, one study found that 150 mg of naltrexone for 12 weeks reduced the daily amount of alcohol consumed, the days engaged in drinking behavior, the amount of money spent per week on cocaine, and the percentage of days using cocaine.¹⁵⁴ However, another study using 50 mg of naltrexone for 12 weeks did not reduce cocaine or alcohol use in people diagnosed with CUD and AUD.¹⁵⁵ Moreover, a third study using 100 mg of naltrexone for 12 weeks found that although rates of cocaine use and drinks per day were not influenced by naltrexone, naltrexone did reduce the frequency of heavy drinking days in people with CUD and AUD.¹⁵⁶ Naltrexone is a competitive μ-opioid receptor antagonist that reduces the reinforcing effects of alcohol by attenuating endogenous opioid-mediated enhancement of mesolimbic dopamine signaling.¹⁵⁷ Given that studies have shown that the reinforcing effects of cocaine can be decreased via opioid antagonists,^158,159 naltrexone may blunt the reinforcing effects of both cocaine and alcohol by reducing opioid-driven facilitation of dopaminergic transmission. However, the mixed findings discussed above suggest that the efficacy of naltrexone to treat co-occurring CUD and AUD may vary depending on factors such as dosage and study endpoints. An important point is that all these clinical trials focused on individuals who met criteria for both CUD and AUD. Yet, it remains unclear whether similar outcomes would be observed in people with CUD who co-use alcohol but do not meet diagnostic criteria for AUD. This highlights the importance of future studies specifically targeting this understudied group, as their treatment needs and responses to medications like naltrexone may differ substantially.

2. Methamphetamine

a. Human laboratory findings

Several studies have investigated how methamphetamine and alcohol interact in humans. In a within-subject study, participants consumed 6 alcoholic drinks for over half an hour, followed by an intravenous injection of methamphetamine administered 60 minutes after the first drink.⁶⁶ The results showed that methamphetamine reduced the subjective effects of alcohol, whereas alcohol did not alter the subjective effects of methamphetamine.⁶⁶ Another study, however, found that when oral methamphetamine was consumed in an alcoholic beverage, the methamphetamine-alcohol combination resulted in greater ratings of “good drug effect” compared with either drug alone.¹⁵⁸ Also, methamphetamine decreased alcohol-induced cognitive and psychomotor impairments. This study also evaluated cigarette smoking and found that participants smoked more cigarettes under the conditions in which alcohol and methamphetamine were co-administered when compared with either drug alone.¹⁶⁰ These findings suggest that methamphetamine and alcohol can alter each other’s subjective effects, but the outcomes may vary depending on the route of administration and the timing of co-use. Moreover, combining alcohol and methamphetamine may also increase the likelihood of self-administering a third substance, such as nicotine.

b. Clinical study findings

To our knowledge, no studies have directly examined pharmacological interventions for MUD while specifically evaluating the role of concurrent alcohol use. In fact, several studies have excluded participants with a history of AUD or controlled for AUD without incorporating it into their analyses.161, 162, 163 Nonetheless, some promising therapeutic targets for individuals who co-use methamphetamine and alcohol include naltrexone and bupropion, as both have been shown in clinical trials, focusing on only one drug, to reduce methamphetamine and alcohol use.164, 165, 166, 167 These potential treatments warrant investigation in populations with co-occurring methamphetamine and alcohol use.

3. Summary

Together, these studies underscore the complex pharmacological interactions and treatment challenges associated with cocaine and alcohol co-use. The combination of these substances can alter both subjective and physiological effects, influence patterns of use and complicate treatment responsiveness. Evidence suggests that co-use may enhance the rewarding and euphoric effects of cocaine while reducing alcohol-induced sedation, contributing to the sustained use of both drugs. Similarly, studies examining methamphetamine and alcohol co-use indicate that these substances interact in ways that depend on the route and timing of administration. Intravenous methamphetamine can dampen the subjective effects of alcohol, whereas oral co-administration with an alcoholic beverage can enhance positive effects and reduce alcohol-related impairments. Co-use may also increase the likelihood of using additional substances, such as nicotine. Overall, interactions between alcohol and cocaine are better characterized, whereas alcohol-methamphetamine interactions remain underexplored in treatment contexts and may depend more on the route and timing of administration.

C. Stimulants and cannabis

1. Cocaine

a. Human laboratory findings

Relatively few studies have directly examined cannabis and cocaine co-use. A double-blind, placebo-controlled study found that smoking a cannabis cigarette 28 minutes before smoking cocaine reduced the typical cocaine-related decrease in hunger and feelings of calm.⁷⁸ Another study reported that smoking cannabis 30 minutes before intranasal cocaine shortened the duration of dysphoria and other negative effects associated with cocaine.⁷⁶ A study where cannabis was smoked 13 minutes before an intravenous injection of cocaine found a trend toward prolonged feelings of stimulation and being high compared with either drug alone.¹⁴⁰ Together, these findings suggest that cannabis may blunt some of cocaine’s negative effects, such as appetite suppression and feelings of agitation, while potentially extending cocaine’s positive subjective effects, like euphoria and stimulation.

b. Clinical study findings

Several studies have explored whether CBD could be a potential treatment for CUD. In one study, individuals with CUD were randomized to receive either 300 mg of CBD or a placebo daily for 10 days. The results showed that cocaine craving declined over time in both groups, with no significant difference between CBD and placebo. This suggests that CBD at that dose had no therapeutic effect.¹⁶⁸ Another trial administered 800 mg of CBD daily for 92 days to individuals with CUD and similarly found no reduction in craving or relapse rates compared with placebo.¹⁶⁹ Furthermore, CBD did not improve cognitive performance across multiple domains in people with CUD.¹⁷⁰ As noted in the ”Stimulants and nicotine” section, cannabis use is generally not an exclusion criterion in clinical trials investigating treatments for CUD. However, individuals may be excluded if they qualify for having a cannabis use disorder.^171,172 Even if people who use cannabis were included, no known studies have examined whether treatment outcomes differ between individuals who co-use cannabis and cocaine versus those who use cocaine alone.

2. Methamphetamine

a. Human laboratory findings

Although the co-use of methamphetamine and cannabis is common among people who use methamphetamine, no human subject studies to date have examined how cannabis use influences the subjective effects of methamphetamine.

b. Clinical study findings

Likewise, no studies have directly compared the effectiveness of treatments for MUD in individuals who co-use cannabis versus those who do not. In fact, many studies either exclude participants with a history of cannabis dependence or allow people who use cannabis to enroll but do not analyze whether treatment outcomes differ by cannabis use status.^162,173 Studies that have considered cannabis use among people with MUD have generally not found significant effects of the interventions on either substance.^174,175 For example, one study found that bupropion was ineffective at reducing methamphetamine or cannabis use among participants who used both substances.¹⁷⁶ However, as with other forms of co-use, no studies, to our knowledge, have stratified participants by cannabis use to evaluate whether treatment efficacy differed between people who did and did not co-use cannabis.

Several clinical studies have investigated how the co-use of cannabis may affect methamphetamine-induced neurocognitive deficits. For example, one study found that individuals with MUD who also used cannabis performed better on measures of verbal fluency, learning, memory, and processing speed than those who used methamphetamine alone.¹⁷⁷ These data suggest that cannabis use did not exacerbate, and may even have mitigated, some methamphetamine-related cognitive impairments.¹⁷⁷ Similarly, another study found that while individuals with MUD exhibited neurocognitive deficits compared with controls, those with a history of cannabis use did not differ significantly from the control group, further supporting that cannabis does not worsen methamphetamine-induced neurocognitive deficits.¹⁷⁸ The mechanisms underlying cannabis's potential to mitigate methamphetamine-related cognitive impairments are unclear. However, cannabinoids have been shown to decrease neuroinflammation and oxidative stress, which may be some mechanisms that counteract the neurotoxic effects associated with chronic methamphetamine use.179, 180, 181 Since chronic stimulant use can result in impairments in cognitive functioning, which may worsen treatment outcomes and daily functioning, clinical studies investigating whether the constituents of cannabis could improve treatment outcomes for MUD are warranted.¹⁸²

3. Summary

For cocaine, limited studies suggest that cannabis may mitigate some of the negative effects of cocaine while potentially enhancing its positive effects. Despite growing interest in CBD as a treatment for CUD, current evidence does not support its efficacy in reducing craving, relapse, or cognitive impairment. For methamphetamine, clinical work has not tested how cannabis influences the acute subjective effects of methamphetamine or the efficacy of treatments. However, some observational studies suggest cannabis does not worsen but may buffer neurocognitive deficits.

D. Stimulants and opioids

1. Cocaine

a. Human laboratory findings

Research examining the subjective effects of co-using cocaine and opioids has shown consistent findings. One study investigated how people experienced the effects of intravenous cocaine and morphine when taken separately or together.¹⁸³ Cocaine on its own led participants to report feeling more “stimulated,” while morphine alone resulted in feelings of “sedation.” When the 2 drugs were administered together in a single injection, participants did not describe any interactions between the 2 substances. One conclusion is that using cocaine and morphine at the same time does not create subjective effects that are noticeably different from using each drug individually.¹⁸³ In a similar experiment, researchers assessed both the physiological and subjective effects of intravenous cocaine, hydromorphone, and their combination in volunteers with a history of cocaine and heroin use.¹⁸⁴ Consistent with the above study,¹⁸³ cocaine and hydromorphone each produced distinct physiological and subjective effects, and the combination generated a profile reflecting individual drug effects rather than an interaction between the drugs. Additionally, the study showed that although naltrexone reduced the subjective and physiological effects of hydromorphone in a dose-dependent manner, it did not alter the subjective and physiological effects of cocaine.¹⁸⁴

Another study evaluated the effects of acute buprenorphine administration in non–opioid-dependent participants on the subjective ratings of cocaine-morphine combinations. Buprenorphine decreased ratings of drug liking and quality.¹⁸⁵ Furthermore, an inpatient study of methadone-maintained individuals with a history of intravenous cocaine and heroin use assessed how methadone and buprenorphine influenced the reinforcing and subjective effects of intravenous cocaine. These investigators found that buprenorphine, but not methadone, reduced cocaine craving. Moreover, when various doses of cocaine could be chosen for over $5, buprenorphine significantly reduced cocaine self-administration while methadone did not.¹⁸⁶ These findings suggest that a partial agonist at μ- and κ-opioid receptors may be more effective in treating the co-use of cocaine and opioids compared with a full μ-opioid agonist.

b. Clinical study findings

Several studies have also been conducted to evaluate pharmacotherapies for individuals with CUD who also use opioids. One study¹⁸⁷ recruited participants with both CUD and co-occurring OUD, and participants were given various sublingual doses of buprenorphine. The researchers assessed drug-positive urine samples along with self-reported use of cocaine and opioids. The results showed that higher doses of buprenorphine significantly reduced opioid-positive urine tests and self-reported opioid use and were also linked to marked decreases in opioid withdrawal symptoms. The impact of buprenorphine on cocaine use was more variable. Overall, there were reductions in cocaine use, but these effects were less pronounced than those on opioid use. Achieving meaningful decreases in cocaine use required higher doses of buprenorphine and longer treatment durations. Participants also reported lower cravings for cocaine during buprenorphine treatment. However, among those who continued using cocaine during the study, ratings of cocaine-induced euphoria did not change based on buprenorphine treatment.¹⁸⁷ Another clinical study evaluating buprenorphine in individuals with co-occurring CUD and OUD found that daily doses of 16 mg buprenorphine resulted in reductions in both opioids and cocaine use, as measured by urine drug screens.¹⁸⁸ However, it is important to note that neither study included participants using only opioids or only cocaine, so it remains unclear whether the treatment effects would differ in the absence of polysubstance use.

One study investigated the effectiveness of disulfiram in treating CUD in methadone-maintained patients with both OUD and CUD, with or without co-occurring alcohol use. The researchers found that participants receiving disulfiram showed a significant reduction in both the frequency and quantity of cocaine use compared with those given a placebo. When alcohol and opioid use were examined over time, all participants reduced their use, regardless of whether they received disulfiram or a placebo. These findings suggest that disulfiram may help reduce cocaine use in individuals with CUD who are maintained on methadone.¹⁸⁹

Another study evaluated the effectiveness of baclofen for the treatment of CUD in participants with CUD alone or CUD and co-occurring OUD in an inpatient setting. The subjects with co-occurring CUD and OUD were maintained on methadone during the study. During the inpatient stay, cocaine self-administration via the smoked route was available several times during baclofen treatments; during these sessions, participants were given the choice to either smoke a sample of cocaine at various doses or receive a merchandise voucher. They found that in the group not dependent on methadone, baclofen decreased cocaine choice compared with the placebo group. Baclofen had no effect on cocaine choice in the methadone-maintained group. These findings suggest that baclofen may only be effective in decreasing cocaine use in non–opioid-dependent patients.¹⁹⁰

A study using a similar cocaine versus merchandise procedure in individuals with CUD only or CUD and OUD evaluated venlafaxine as a potential treatment for CUD.¹⁹¹ Participants with OUD were maintained on methadone throughout the study. They found that while venlafaxine decreased the subjective effects of cocaine in people with CUD alone and those with CUD and OUD, it had no effect on cocaine choice in either group.¹⁹¹ These results highlight that the effectiveness of pharmacotherapies for CUD can differ between people with CUD alone and those with CUD and OUD, and that these differences may depend on the medication tested. Moreover, the findings suggest that in clinical studies, there may be a disconnect between subjective reports and reinforcing effects.

2. Methamphetamine

a. Clinical study findings

While human-subject studies have not investigated whether opioids impact the subjective effects of methamphetamine, a recent systematic review¹⁹² examined studies investigating the relationship between methamphetamine or amphetamine use and 3 key outcomes related to OUD: receipt of medication for OUD; retention in treatment with a medication for OUD; and opioid abstinence during treatment. These findings will be briefly summarized. In general, most studies,193, 194, 195, 196, 197 although not all,198, 199, 200 found that methamphetamine use was negatively associated with receiving medication for OUD, particularly methadone or buprenorphine. In short, individuals who used methamphetamine were less likely to receive these medications despite using opioids. Moreover, most,^199,201, 202, 203, 204, 205, 206, 207, 208 but not all,209, 210, 211 studies found a negative association between retention in treatment for an OUD and methamphetamine use. Regarding opioid abstinence, about half of the studies212, 213, 214, 215 found a negative association between methamphetamine use and opioid abstinence, while the other half216, 217, 218, 219 found no significant association. These findings suggest that methamphetamine use can, under certain conditions, reduce the likelihood of receiving medication for OUD, lower treatment retention, and decrease the chances of maintaining opioid abstinence.

3. Summary

In summary, data on the subjective effects of cocaine and opioids suggest that co-use does not produce amplifications of either drug’s effects. Moreover, the findings from clinical trials examining novel therapeutics for CUD raise important questions about the role of opioid dependence and maintenance therapy in treating CUD among people who also use opioids. Data suggest that individuals who use methamphetamine, concurrent opioid use is linked to lower rates of receiving or remaining on medication for OUD and, in some cases, reduced opioid abstinence during treatment. Together, these findings suggest that opioids complicate care across both CUD and MUD by influencing treatment access, retention, and pharmacological responsiveness. This underscores the importance of accounting for opioid co-use when assessing stimulant-related outcomes.

E. Main takeaways

The clinical data reviewed here (summarized in Table 2) suggest that the reinforcement mechanisms (positive and negative) outlined earlier, underlie many of the observed co-use patterns of cocaine and methamphetamine. For example, in human laboratory studies, cocaine-alcohol combinations consistently produced heightened euphoria, which likely perpetuates continued co-use (positive reinforcement). Likewise, cannabis reduced cocaine-related dysphoria and agitation, which likely also perpetuates continued co-use patterns (negative reinforcement). These 2 examples, along with the others described above, illustrate how positive and negative reinforcement mechanisms maintain stimulant polysubstance use. Future clinical studies that explicitly measure motivational drivers alongside pharmacological outcomes could help clarify these pathways and improve treatment targeting.

Table 2.

Summary of human laboratory and clinical findings on polysubstance use involving cocaine and methamphetamine

Co-Used Substance	Cocaine-Related Findings	Methamphetamine-Related Findings
Nicotine	Nicotine heightened cocaine cravings, the amount of cocaine used, and smoking cessation improved treatment outcomes for CUD.120, 121, 122, 123,125	Reduced smoking did not impact methamphetamine use.131,132
Alcohol	Alcohol enhanced the pleasurable subjective effects of cocaine, while cocaine reduced the negative effects of alcohol, like sedation.61,135, 136, 137, 138, 139, 140, 141 Co-use reduced the efficacy of treatments for CUD.127,143,149	Mixed findings, suggesting that alcohol can either have no effect or increase the positive subjective effects of methamphetamine.158,160 Methamphetamine also decreased alcohol-induced cognitive and motor impairments.66,160
Cannabis	Cannabis blunted the negative effects of cocaine, like appetite suppression, while extending positive subjective effects.76,78,140 CBD is not effective for treating CUD.168, 169, 170	Cannabis may mitigate methamphetamine-related cognitive deficits.177,178
Opioids	Co-use does not amplify drug effects, and the effect of co-use in opioid-dependent individuals may vary based on the therapeutic tested.183, 184, 185, 186, 187, 188, 189, 190, 191	Opioid use is linked to lower rates of receiving or remaining on medication for OUD and reduced opioid abstinence during treatment.193, 194, 195, 196, 197,199,201, 202, 203, 204, 205, 206, 207, 208,212, 213, 214, 215

However, a recurring limitation is that very few clinical trials separate participants into groups based on whether they use only a single drug (eg, cocaine alone) or a drug combination (eg, cocaine and cannabis). This might be because recruiting participants who use only one stimulant, in the absence of co-use of other substances, is extremely challenging. For instance, if 85% of people who use cocaine smoke cannabis,²⁸ enrolling a large enough group who use cocaine only would be very difficult. This becomes even more complicated when considering that many people co-use 2 or more additional substances alongside stimulants. This makes it impractical to design clinical studies with sufficient numbers of participants for each drug combination. Given these constraints, many of these questions may be better addressed initially through preclinical research. Animal models offer more precise control over drug exposure and history and allow researchers to test multiple groups cost effectively (especially when using rodents). These models can help build a clearer foundation for understanding how polysubstance use shapes the behavioral effects of stimulants and treatment outcomes. The role of preclinical models in addressing these questions is discussed next.

V. Preclinical findings of stimulant polysubstance use

Given the difficulty of drawing causal inferences from human epidemiological research and the inherent limitations of human clinical studies, preclinical studies using animal models provide a critical complement to advancing our understanding of polysubstance use involving stimulants. Unlike human studies, animal research allows for precise experimental control over drug exposure, the environment, and the time course of drug co-use, making it possible to systematically examine causal mechanisms underlying polysubstance use with stimulants.²²⁰ These animal models offer unique advantages by enabling researchers to examine the entire substance use cycle, including factors that modulate vulnerability to substance use, the maintenance of drug self-administration, and the efficacy of behavioral and/or pharmacological interventions for substance use.²²¹ Thus, preclinical animal studies can help disentangle the directionality of effects that modify the likelihood of polysubstance use, isolate specific biological and behavioral mechanisms, and evaluate pharmacological and behavioral interventions with high translational relevance. The following section begins with a brief overview of the animal models used to study drug use. We will then highlight key findings from behavioral preclinical studies that examined stimulant polysubstance use involving cocaine and methamphetamine across various stages of the substance use cycle.

A. Animal models of drug reinforcement

Several well-established procedures are used in preclinical research to study the behavioral and motivational effects of drugs in animal models. Among the most widely employed techniques are self-administration paradigms, which directly assess a drug’s reinforcing effects, and related methods such as drug discrimination, reinstatement, and conditioned place preference (CPP), which evaluate subjective, relapse-related, or conditioned rewarding effects, respectively. Although there are several other procedures, including the assessment of unconditioned behaviors that are used to examine the effects of drugs in animal models, these paradigms will not be discussed here. However, additional details about these procedures can be found elsewhere.222, 223, 224

Preclinically, one of the most widely used animal models is drug self-administration.225, 226, 227 These studies involve training an animal to respond on an operandum (eg, lever or breaking a photo-optic switch), in which a stimulus (eg, light) signals that a drug will be delivered after a behavioral response (eg, lever press or finger poke). During drug self-administration, responses produce stimulus changes (eg, drug delivery) according to rules that specify the contingency or schedule of reinforcement. These procedures can be used to examine how polysubstance use influences the reinforcing potency and/or strength of drugs under various schedules of reinforcement, including ratio- and interval-based schedules.

Although beyond the scope of this review, the schedule of reinforcement used in drug self-administration studies should depend on the questions the investigator is asking. For example, if one were interested in questions related to sensitivity to function as a reinforcer, simple schedules of reinforcement, such as fixed-ratio (FR) schedules, in which a specific number of responses is required for reinforcement. Adding a second drug could increase the potency of a stimulant (eg, cocaine) to function as a reinforcer. If the investigator was interested in questions about whether polysubstance use leads to stronger reinforcing effects compared with either drug alone, progressive-ratio (PR) schedules of reinforcement would be better suited. In PR schedules, the response requirement for a drug injection increases until a “breakpoint” is reached (the maximum number of injections in a session). If treatments were to be examined, concurrent choice paradigms between drug and nondrug alternatives would be ideal.

All drugs produce unique internal or subjective effects. In animal models, these effects can function as discriminative stimuli, allowing researchers to train animals to distinguish between a drug and its vehicle (or between different drug classes).²²⁸ In the drug discrimination procedure, a dose of drug or vehicle is non-contingently administered, and the animal is then trained to respond to an operant, as described for self-administration procedures. During these operant sessions, animals generally have access to 2 manipulanda (eg, 2 levers) for behavioral responses. Behavioral responses (ie, lever presses) to the drug-designated lever produce reinforcement (ie, food pellet) only during sessions in which the training drug was administered; during sessions in which saline was administered, lever presses on the saline-designated lever would produce reinforcement.²²⁹ Once the animal is trained on this procedure, alterations in a training drug’s discriminative stimulus effects can be tested under various conditions, including those in which other drugs are co-administered.

In addition, studies using animal models of drug self-administration have examined how variables like drug, drug-associated cues, and exposure to certain stressors can increase drug-related behaviors following extinction. Extinction, in this context, is defined as a significant reduction in drug-associated behaviors when behavioral responses no longer result in drug delivery.²³⁰ Broadly, the reinstatement model is thought to measure how particular variables influence susceptibility to relapse to drug use after abstinence.²³¹ Thus, one variable that can be assessed using this method is how the co-administration of drugs influence reinstatement compared with the administration of a drug in isolation.

Another procedure commonly used preclinically is the CPP paradigm. This is often cited as measuring the “rewarding” properties of drugs through principles related to Pavlovian conditioning.^224,232 This procedure uses a 2-compartment box with distinct sensory cues. Animals receive a drug or vehicle injection and are placed in a specific compartment, repeatedly pairing drug effects and vehicle with that environment. In a later drug-free test, animals explore both compartments freely, and preference is measured by the time spent in each. Studies implementing this procedure using numerous drugs have found that when a drug of abuse was paired with a particular compartment, animals would spend significantly more time in that compartment compared with the non–drug-associated compartment during test sessions in the absence of an injection.233, 234, 235 One interpretation of the increase in preference for the drug-paired compartment following conditioning is that the animal learned to associate the rewarding effects of the drug stimulus with the cues in that compartment. Importantly, CPP paradigms can be used to assess polysubstance use by examining whether co-administration of drugs increases or decreases CPP.

Collectively, these procedures—self-administration, drug discrimination, reinstatement, and CPP—form the foundation of preclinical research on measures of abuse liability and continue to be instrumental in advancing our understanding of substance use. Notably, many animal studies investigating polysubstance use, particularly those involving stimulants such as cocaine and methamphetamine, rely on these well-established paradigms to assess how combinations of drugs influence reinforcing potency, reinforcing strength, subjective effects, reinstatement, and conditioned reward.

Before describing specific preclinical studies, it is important to address a critical methodological issue: how the drugs are administered. Are all drugs administered contingently (self-administered by the animal), non-contingently (experimenter-administered), or is one drug self-administered while others are administered non-contingently? Is the co-used drug given simultaneously with the stimulant, or does its administration precede or follow stimulant exposure, and how often are these drugs delivered in a session? Many studies favor non-contingent administration of multiple substances because it offers greater experimental control over dosing and timing. However, this approach can limit the translational relevance of the findings. Striking the right balance is crucial, and at least one review has argued that the field should move away from an excessive focus on experimental “purity” to better reflect the complex realities of real-world polysubstance use.²³⁶ More broadly, there is a growing consensus that additional research on polysubstance use is needed across the entire translational research pipeline, including preclinical models, to better capture and address real-life patterns of co-use.

B. Stimulants and nicotine

1. Cocaine

Numerous preclinical studies have examined how nicotine use influences the rewarding and reinforcing effects of cocaine. Experiments using a CPP procedure have shown that nicotine increased the rewarding effects of cocaine.²³⁷ For example, one study found that a single subcutaneous injection of nicotine increased the time spent in a cocaine-paired chamber.²³⁷ Similarly, one study examining the discriminative stimulus effects of cocaine demonstrated that nicotine potentiated these effects as well. In monkeys, acute nicotine administration increased the discriminative stimulus effects of low cocaine doses and did not substitute for cocaine.²³⁸ This suggests that nicotine may increase the subjective effects of low doses of cocaine in humans. In terms of the effect of nicotine on cocaine reinforcement, a study using drug-naïve male rats examined the acquisition of cocaine reinforcement at a single dose (0.25 mg/kg per injection) on an FR 1 schedule of reinforcement.²³⁹ They found that 9 days of pre-exposure to subcutaneous nicotine (0.6 mg/kg) increased response rates during the early days of cocaine acquisition compared with control rats. In that study, acquisition was examined 24 hours after the last nicotine pretreatment.²³⁹ These data suggest that nicotine increased the reinforcing effects of cocaine, even when the administration of the 2 drugs was temporally separated.

Another study of male rats evaluated the effect of a subcutaneous injection of nicotine (0.15, 0.3, and 0.6 mg/kg) 3 minutes prior to cocaine self-administration, at a dose on the ascending limb of the dose-response curve, on a PR schedule of reinforcement (ie, the number of responses needed to deliver each subsequent cocaine injection was progressively increased).²⁴⁰ They found that acute nicotine pretreatment increased the number of cocaine infusions at a low nicotine dose but decreased the number of cocaine infusions at a higher nicotine dose. However, when the high nicotine dose was given chronically, nicotine administration increased the number of cocaine infusions by day 8 of treatment. They also found that after chronic treatment with subcutaneous nicotine injections prior to cocaine self-administration, nicotine injections reinstated responding for cocaine after several days of extinction training under the PR contingencies.²⁴⁰ However, one limitation of the above studies is that only one dose of cocaine was tested with nicotine; thus, it is not possible to determine whether increased response rates after nicotine administration reflected leftward or rightward shifts in the cocaine dose-response curve.

Other investigations examining cocaine and nicotine co-use have studied a range of cocaine doses and found that nicotine appeared to potentiate cocaine reinforcement. One study found that on a PR schedule of reinforcement, adding nicotine (0.03 mg/kg) to the cocaine drug bag at a variety of doses resulted in an upward shift in the cocaine dose-response curve in male rodents, suggesting an increase in the reinforcing strength of cocaine with the addition of nicotine.²⁴¹ However, another study of male rhesus monkeys found that on a PR schedule of reinforcement, adding nicotine (0.012–0.05 mg/kg) to the cocaine drug bag resulted in leftward shifts in the cocaine-dose response curve but not upward shifts.²⁴² Thus, nicotine increased the potency of cocaine to function as a reinforcer, but not its reinforcing strength, in monkeys.

When the above study²⁴² in rhesus monkeys was recently replicated by another laboratory, adding nicotine (0.01–0.03 mg/kg) to the cocaine drug bag resulted in upward and leftward shifts in the cocaine dose-response curves on a PR schedule of reinforcement.²⁴³ As a result, it is likely that methodological differences account for whether nicotine simply increased the potency of cocaine or the reinforcing strength. The latter study²⁴³ also evaluated cocaine and nicotine co-use under a concurrent drug-versus-food choice schedule (behavior must be allocated to one of 2 manipulanda for the delivery of drug or food pellets) in male and female socially housed cynomolgus monkeys. They found that nicotine shifted the cocaine dose-effect curve to the left in 13 of 14 monkeys, but the potentiation of cocaine choice with the addition of nicotine was greater in female monkeys than in male monkeys. Furthermore, when a behavioral intervention was implemented (ie, delay discounting), both male and female monkeys were willing to wait significantly longer for an injection of cocaine when it was co-administered with nicotine compared with cocaine alone,²⁴³ supporting the hypothesis that nicotine increased the reinforcing strength of cocaine.

These findings suggest that nicotine increases the discriminative stimulus effects of cocaine and induces a stronger CPP compared with cocaine alone. Also, under certain experimental conditions, nicotine not only increases the potency of cocaine but also the reinforcing strength of cocaine. Nicotine may also reduce the efficacy of behavioral interventions aimed at reducing cocaine self-administration. Moreover, results from one study suggest that chronic nicotine exposure may play a role in the reinstatement of extinguished cocaine-maintained responding. The effect of nicotine on cocaine is consistent with positive reinforcement driving co-use patterns, as co-use may increase reinforcing strength and potency.²⁴⁰

2. Methamphetamine

To date, several preclinical studies have examined the co-use of nicotine with methamphetamine. However, most of these studies have focused on the effects of chronic nicotine exposure on methamphetamine reinforcement rather than investigating the acute interactions between the 2 substances when administered in close succession or simultaneously. For example, a study of male rats using an FR schedule showed that administering a low dose of nicotine daily for 15 days during adolescence (PND 35–50) led to increased methamphetamine intake in adulthood at the methamphetamine dose tested.²⁴⁴ However, nicotine exposure did not influence methamphetamine-primed reinstatement. Another study using male and female rats found that exposure to nicotine during early adolescence increased oral methamphetamine self-administration on an FR schedule of reinforcement, particularly in female rats.²⁴⁵ However, if nicotine exposure occurred later in adolescence, nicotine decreased oral methamphetamine self-administration, independent of sex. Although neither early nor late nicotine exposure affected the rate of extinction once oral methamphetamine was no longer available following completion of the FR requirement, early nicotine exposure resulted in lower responding during reinstatement in male rats after methamphetamine priming.²⁴⁵ Moreover, consistent with sex differences in the interactions between nicotine and methamphetamine, another study found that a history of nicotine exposure selectively increased methamphetamine self-administration in female rats but not in male rats.²⁴⁶

An additional study examined the effects of a brief 4-day nicotine exposure beginning on either PND 28 (adolescence) or PND 86 (adulthood). Under an FR schedule and a single methamphetamine dose, male rats exposed to nicotine during adolescence (PND 28) self-administered more methamphetamine compared with age-matched controls that received saline. This effect was not observed in males exposed to nicotine during adulthood (PND 86), suggesting an age-dependent sensitivity to the effects of nicotine on methamphetamine reinforcement. In female rats, the effects of adolescent nicotine exposure were not evident; however, nicotine-exposed adult females self-administered more methamphetamine than nicotine-exposed adolescent females.²⁴⁷ In another preclinical study, researchers investigated how prenatal exposure to nicotine influenced methamphetamine self-administration at one dose on a PR schedule of reinforcement in adult rat offspring. While sex differences were not significant, prenatally nicotine-exposed rats had significantly higher breakpoints than controls, suggesting that prenatal nicotine exposure increased the reinforcing strength of methamphetamine.²⁴⁸ Overall, these studies indicate a potential sex-specific interaction between age of nicotine exposure and methamphetamine self-administration and reinstatement. Furthermore, these data suggest that prenatal nicotine exposure contributes to increased methamphetamine reinforcement.

A study investigating both acute and chronic nicotine exposure on methamphetamine reinforcement in rats found that acute nicotine administration reduced methamphetamine self-administration, but only at the highest nicotine dose tested (0.05 mg/kg); an effect likely driven by potentiating the rate-decreasing effects of methamphetamine rather than a true decrease in drug reinforcement.²⁴⁹ In contrast, repeated administration of a lower nicotine dose over 14 days had no effect on methamphetamine intake. Following extinction, a nicotine pretreatment significantly reinstated extinguished methamphetamine responses, but only in rats that had received chronic nicotine exposure over the 14-day period. Rats that received only saline during this phase did not show an increase in extinguished responding following a nicotine pretreatment.²⁴⁹

While the above study²⁴⁹ emphasized the importance of prior nicotine exposure in facilitating methamphetamine reinstatement, another study²⁵⁰ suggested that acute nicotine can increase the effects of methamphetamine, even in the absence of prior nicotine exposure, using a CPP paradigm. In adult male rats, nicotine administered for 7 days prior to methamphetamine conditioning had no effect on the acquisition, rates of response extinction, or reinstatement of methamphetamine CPP. However, when nicotine (0.2 mg/kg) was administered acutely in combination with a dose of methamphetamine (0.5 mg/kg) during the reinstatement phase, it successfully reinstated CPP, even though this methamphetamine dose was ineffective on its own.²⁵⁰ One potential mechanism underlying the acute interactions between nicotine and methamphetamine is their shared discriminative stimulus effects.^251,252 Supporting this premise, a study of rats trained to discriminate nicotine found that methamphetamine produced full substitution, albeit with a lower potency. Conversely, in rats trained to discriminate methamphetamine, nicotine produced partial substitution but with greater potency than methamphetamine.²⁵³ These findings suggest that nicotine may acutely increase the interoceptive effects of methamphetamine, which may contribute to co-use, even in the absence of chronic prior nicotine exposure.

Several studies suggest that nicotine may mitigate certain cognitive deficits associated with methamphetamine use.²⁵⁴ For example, in one study, rats treated with methamphetamine (4.0 mg/kg) for 7 days exhibited lasting impairments in spatial working memory, which persisted for at least 2 weeks after drug cessation.²⁵⁵ However, when nicotine (0.3 mg/kg) was administered daily during the abstinent period, these cognitive impairments were significantly attenuated in the first week. However, this effect was diminished by the second week of drug cessation, indicating that nicotine may have some short-lasting effects on methamphetamine-related working memory deficits.²⁵⁵ Similarly, another study found that 30 days of methamphetamine treatment (4.0 mg/kg) impaired decision-making in rats in a task assessing high-risk behavior.²⁵⁶ However, this impairment was not observed in rats that received concurrent daily nicotine treatment (0.3 mg/kg), suggesting a protective effect of nicotine. Notably, nicotine treatment alone did not affect decision-making in methamphetamine-naïve rats.²⁵⁶ Together, these findings suggest that nicotine may offer limited and transient protective effects against some of the cognitive impairments induced by chronic methamphetamine exposure.

Preclinical research examining nicotine and methamphetamine co-use reveals a relationship shaped by age, sex, timing, and pattern of nicotine exposure. Chronic nicotine exposure, particularly during adolescence, often enhances methamphetamine self-administration, with female rats showing heightened sensitivity. Prenatal nicotine exposure also increases methamphetamine’s reinforcing strength in adult offspring. While acute nicotine occasionally reduces methamphetamine intake at high doses, likely due to nonspecific rate-decreasing effects, studies found that repeated nicotine potentiates methamphetamine-primed reinstatement. Furthermore, even in the absence of prior nicotine exposure, acute nicotine could reinstate methamphetamine-induced CPP when combined with subthreshold methamphetamine doses, potentially due to shared discriminative stimulus effects. In addition to altering reinforcement, nicotine may transiently mitigate methamphetamine-induced cognitive impairments, such as deficits in working memory and decision-making. Many of these findings are consistent with positive reinforcement driving co-use, where nicotine enhances the reinforcing or interoceptive effects of methamphetamine. However, it remains unclear whether nicotine might also be used to offset the aversive or withdrawal-related effects of methamphetamine. Collectively, these findings suggest that nicotine can both exacerbate methamphetamine use and, under certain conditions, attenuate its cognitive consequences.

3. Summary

Across species and experimental procedures, nicotine consistently increases cocaine reinforcement and potentiates both the discriminative stimulus and CPP effects of cocaine. Nicotine can also interfere with behavioral interventions aimed at reducing cocaine use. In contrast, interactions between nicotine and methamphetamine are more context-dependent and vary by exposure history, sex, and developmental timing. Overall, nicotine reliably increases cocaine reinforcement, whereas its effects on methamphetamine are more variable but still tend to increase methamphetamine use.

However, it is important to note that these observed effects do not necessarily imply a unidirectional enhancement of stimulant reinforcement by nicotine. Many of the outcomes described may instead reflect additive or bidirectional interactions between nicotine and the stimulant. Consequently, the apparent potentiation of stimulant reinforcement could arise from nicotine amplifying responses to drug-paired cues rather than directly altering the pharmacological reinforcing strength of the stimulant itself. Determining the directionality and mechanism of these interactions will require comprehensive dose-response studies of both drugs and paradigms designed to disentangle the pharmacological, conditioned, and behavioral components of co-use. As will become evident across subsequent preclinical sections, clarifying these relationships is essential for interpreting stimulant polysubstance interactions with greater precision.

C. Stimulants and alcohol

1. Cocaine

Preclinical research using rodents and monkeys has also examined how ethanol influences cocaine reinforcement. Studies examining whether ethanol exposure prior to cocaine self-administration increased vulnerability to cocaine reinforcement have found mixed results. One experiment found that male rats consuming higher levels of ethanol over 36 sessions consumed more cocaine during initial exposure to cocaine.²⁵⁷ Similarly, another study observed that both male and female mice receiving intraperitoneal ethanol injections over a 2-week period had a higher number of cocaine infusions.²⁵⁸ However, these studies tested only a single cocaine dose, limiting conclusions about whether the increased intake reflected heightened sensitivity to cocaine reinforcement. In contrast, an investigation using an escalating dose approach to examine cocaine reinforcement following intermittent ethanol self-administration over 7 weeks in male rats found no link between prior ethanol use and later cocaine self-administration.²⁵⁹ Moreover, a study of male rhesus monkeys showed that even after 9 months of voluntary ethanol intake up to 2.0 g/kg, there were no significant differences in cocaine acquisition compared with ethanol-naïve controls, and ethanol consumption levels did not predict susceptibility to cocaine reinforcement.²⁶⁰

Notably, in each of the 3 rodent studies mentioned above,257, 258, 259 there was a minimum 2-week gap between the period of ethanol exposure and the start of cocaine self-administration. Additionally, in the nonhuman primate study,²⁶⁰ ethanol and cocaine access were spaced at least 3 hours apart to attenuate any overlap in their PD or PK effects. However, as mentioned in previous sections, studies suggest that the PD and PK interactions between alcohol and cocaine may play a critical role in maintaining co-use.³⁶ However, to our knowledge, only one study has evaluated how ethanol self-administration, when administered sequentially to cocaine self-administration, influenced vulnerability to cocaine reinforcement.¹³⁰ In that experiment, the acquisition of cocaine self-administration was studied in a group of same-sex socially housed female and male monkeys that not only had a chronic history of ethanol self-administration prior to exposure to cocaine but also self-administered ethanol immediately prior to cocaine self-administration. Previous research on non–alcohol-drinking monkeys has shown that subordinate monkeys were more sensitive to the reinforcing effects of cocaine than dominant monkeys.^261,262 In contrast to what was hypothesized, these investigators found that oral ethanol consumption decreased vulnerability to cocaine reinforcement but only in subordinate monkeys of both sexes.²⁶³ Subordinate monkeys are thought to be chronically stressed,²⁶⁴ typically lose all fights, are groomed the least, and have the least access to food and treats. One hypothesis for these results was that ethanol, due to its acute anxiolytic effects, reversed the effects of chronic stress in subordinate monkeys and thereby blunted sensitivity to cocaine reinforcement during the initial exposure to cocaine.

The results of studies examining how ethanol influences cocaine use beyond initial vulnerability have also been mixed. One study using CPP in rats found that ethanol could both weaken and strengthen cocaine-induced CPP, depending on the dose of cocaine given during conditioning.²⁶⁵ When high doses of cocaine were used in conjunction with ethanol, ethanol attenuated cocaine-induced CPP; however, when lower doses of cocaine were given with ethanol, ethanol strengthened cocaine-induced CPP. This demonstrates that the effect of ethanol on cocaine CPP was dependent on the dose of cocaine and cannot be explained in terms of augmenting or attenuating cocaine potency.²⁶⁵ Another study found that in rodents, pretreatment with non-contingent intravenous cocaine immediately prior to being given access to an ethanol solution in the home cage resulted in greater ethanol self-administration compared with control animals.²⁶⁶

However, studies on nonhuman primates have not been consistent. One study found that non-contingent administration of intravenous ethanol (0.1–1.7 g/kg) 10 minutes prior to cocaine had no significant effect on cocaine-maintained responding on an FR schedule of reinforcement.²⁶⁷ In another study using non-contingent intravenous administration of ethanol immediately prior to cocaine self-administration on a fixed-interval (FI) 300-second schedule of reinforcement and low doses of cocaine (0.001–0.01 mg/kg), 0.25 and 0.5 g/kg of ethanol did not alter cocaine self-administration, whereas administration of 1.0 g/kg of ethanol decreased the number of cocaine injections from an average of 2.9 to 1.1.²⁶⁸ At high cocaine doses (0.1 mg/kg), no ethanol dose had an effect on the number of cocaine injections.²⁶⁸

Moreover, another study using non-contingent oral ethanol dosing (0.5–2.0 g/kg) 30 minutes prior to cocaine self-administration on a second-order FI schedule of reinforcement found that ethanol dose-dependently decreased cocaine self-administration in 3 of the 4 monkeys.²⁶⁹ Finally, a recent study using socially housed male and female monkeys demonstrated that when ethanol was self-administered immediately prior to cocaine self-administration on an FR schedule of reinforcement, ethanol dose-dependently decreased the potency of cocaine to function as a reinforcer, regardless of sex or social rank.²⁷⁰ Moreover, studies using drug discrimination evaluating the effects of ethanol on cocaine discrimination in rats have shown that ethanol fully blocked the discriminative stimulus effects of low doses of cocaine.²⁷¹

To our knowledge, only one study²⁷² has examined how the co-use of ethanol and cocaine in animal models may influence the efficacy of behavioral and pharmacological interventions. In that study,²⁷² socially housed male and female cynomolgus monkeys self-administered ethanol (1.5 g/kg) for 1 hour prior to a concurrent cocaine-versus-food choice procedure. Ethanol self-administration did not alter the potency of cocaine, but when delays were imposed on access to the lowest preferred cocaine dose, monkeys required longer delays to shift their response toward food following ethanol self-administration. This indicates that ethanol increased the reinforcing strength of cocaine and diminished the effectiveness of interventions designed to shift the choice to non-drug alternatives. This is an important area for further research, and notably, no preclinical studies have systematically evaluated candidate pharmacotherapies for CUD under conditions of concurrent ethanol and cocaine use.

As introduced in earlier sections, the psychoactive metabolite cocaethylene may also play a role in ethanol and cocaine co-use. Studies have shown that intravenous cocaethylene shares discriminative stimulus effects with cocaine and is self-administered with equal potency to cocaine in PR and concurrent choice procedures in monkeys.^68,71,273 However, because the PK profiles of cocaethylene differ between intravenous administration and formation in vivo, it is still unclear how cocaethylene modulates cocaine and ethanol co-use.^136,274 This is also an important future direction for preclinical studies examining cocaine and alcohol co-use.

2. Methamphetamine

A limited number of preclinical behavioral studies have explored the interactions between methamphetamine and alcohol. Although several rodent studies suggest that ethanol may potentiate methamphetamine-induced neurotoxicity, findings examining the behavioral interactions between methamphetamine and alcohol have been mixed.^275,276 One of these studies found that mice with a 10-day history of consuming ethanol in a home-cage binge drinking paradigm exhibited lower oral methamphetamine intake across a range of concentrations on an FR schedule compared with mice with no prior history of ethanol consumption.²⁷⁷ Another experiment examined how a prior history of oral methamphetamine consumption (10 days) influenced home-cage ethanol binge drinking and found that methamphetamine-experienced mice exhibited a greater total daily intake of ethanol compared with controls, particularly at higher concentrations of ethanol (40%).²⁷⁷

Next, these investigators examined what would happen if the mice were given a choice between oral ethanol alone, methamphetamine alone, and a mixture of ethanol and methamphetamine.²⁷⁷ Drug-naïve controls showed a greater preference for the mixture over either single drug solution. In contrast, mice with an ethanol binge history only and those with a history of consuming both ethanol and methamphetamine drank equal amounts of each solution. These findings demonstrate the importance of prior drug history in shaping subsequent patterns of methamphetamine and alcohol co-use. Specifically, a history of binge drinking ethanol reduced methamphetamine self-administration, while prior methamphetamine exposure increased subsequent ethanol intake. Additionally, the experiment suggests that a combination of alcohol and methamphetamine may be particularly appealing to less drug-experienced individuals.²⁷⁷ However, a separate study of rats revealed a different pattern: 21 days of ethanol access using a 2-bottle choice procedure had no impact on later intravenous methamphetamine (0.1 mg/kg) self-administration on an FR schedule, but significantly reduced methamphetamine-primed reinstatement after extinction.²⁷⁸

A third study²⁷⁹ examined concurrent access to ethanol and methamphetamine. After 1 week of access to a 2-bottle choice procedure (10% ethanol versus water) in the home cage, rats were trained to self-administer intravenous methamphetamine (0.1 mg/kg) in an operant chamber. Throughout methamphetamine self-administration, the ethanol 2-bottle choice procedure remained available in the home cage. Under these conditions, rats that consumed ethanol had higher rates of methamphetamine self-administration on an FR 1 schedule during initial exposure compared with controls. However, this effect did not persist with more demanding schedules, including FR 5 and PR schedules. Interestingly, ethanol preference in the home-cage progressively decreased on days when methamphetamine was available, relative to pre-methamphetamine levels. After cessation of methamphetamine self-administration, ethanol intake rebounded to pre-methamphetamine levels.²⁷⁹ These findings suggest that concurrent ethanol availability may initially increase methamphetamine intake under low effort conditions. However, this effect does not persist under higher-effort conditions. Moreover, in this study, methamphetamine availability suppressed ethanol preference.²⁷⁹

These studies reveal varying interactions between methamphetamine and alcohol, which are shaped by factors such as drug history, access timing, and the contingencies of drug administration. Methodological differences likely account for many of the discrepancies observed across studies. For example, the route of methamphetamine administration (oral versus intravenous), the ethanol intake paradigm (eg, binge drinking versus 2-bottle choice continuous access), and the temporal relationship between ethanol and methamphetamine access vary across experiments. Additionally, differences in reinforcement schedules can influence observed interactions between methamphetamine and alcohol.

3. Summary

The combination of these studies suggests that ethanol, under certain conditions, can increase, decrease, or have no effect on the potency of cocaine to function as a reinforcer. In addition, ethanol may block the subjective effects of low doses of cocaine and have varying effects on CPP, depending on the dose of cocaine. Moreover, the metabolite cocaethylene may also play a role in the effect that ethanol has on cocaine reinforcement. Some preclinical findings are consistent with cocaine and alcohol co-use patterns being driven by positive reinforcement. However, other studies show no effect or even attenuating effects of ethanol on cocaine reinforcement, and these findings do not support a reward enhancement interpretation. Negative reinforcement mechanisms, such as co-use to relieve anxiety, dysphoria, or withdrawal from cocaine, are difficult to model rigorously in animal procedures and, to our knowledge, have not been systematically tested in ethanol-cocaine paradigms.

Regarding methamphetamine co-use with alcohol, some findings, such as increased methamphetamine intake following ethanol consumption, are also consistent with positive reinforcement maintaining co-use. In contrast, other results, including reduced reinstatement after prior ethanol exposure, do not support this explanation. Similar to cocaine and alcohol findings, no preclinical studies have rigorously examined whether methamphetamine and ethanol co-use might be maintained by negative reinforcement mechanisms. Overall, these results highlight the complexity of modeling polysubstance use in preclinical settings and underscore the need to carefully consider experimental design when interpreting behavioral outcomes.

D. Stimulants and cannabis

1. Cocaine

Several preclinical studies have examined how the main constituents of cannabis, THC and CBD, influence cocaine use. A rodent study²⁸⁰ found that escalating doses of intraperitoneal injections of THC twice daily (2–8 mg/kg/injection) for 3 days had no effect on the acquisition of cocaine reinforcement on an FR schedule of reinforcement at one dose of cocaine (0.3 mg/kg/injection). After acquisition of cocaine reinforcement, various doses of cocaine were tested, and again, there was no effect of THC exposure on the potency of cocaine to function as a reinforcer. When the contingency was switched to a PR schedule, THC administration resulted in a downward shift in the cocaine dose-response curve, suggesting a decrease in the reinforcing strength of cocaine.²⁸⁰ Another study found that adolescent rats pretreated with 1.0 mg/kg THC i.p. for 17 days had an increase in the potency of cocaine to function as a reinforcer in adulthood when responding under an FR 5 schedule of reinforcement; when the rats were transitioned to a PR schedule of reinforcement, there was no change in the reinforcing strength of cocaine based on THC exposure during adolescence.²⁸¹ It is important to note that in the above studies, cocaine self-administration did not occur in the presence of acute THC administration but rather after repeated THC exposure.

Further studies evaluating the acute interactions between CBD/THC and cocaine found that acute CBD had no effect on cocaine self-administration at one dose (0.5 mg/kg/injection) in rats responding to an FI 20-second schedule of reinforcement and on a PR schedule of reinforcement. Moreover, CBD did not influence cue-induced reinstatement of extinguished cocaine self-administration when given acutely prior to the session.²⁸² To date, only one study has examined the effects of acute pretreatments of THC (0.03–0.3 mg/kg) on cocaine choice in monkeys and found that THC increased cocaine choice at doses that were normally not self-administered over the food alternative reinforcer.²⁸³ Finally, in a study evaluating CPP and cocaine in rats, acute pretreatment of either THC or CBD resulted in a decrease in cocaine-induced CPP compared with rodents given a vehicle pretreatment.²⁸⁴ No preclinical studies are currently available that have evaluated how acute CDB or THC exposure influences the subjective interoceptive effects of cocaine in drug discrimination paradigms. This is an important area for future study.

These findings suggest a divergent interaction between cannabis constituents and cocaine use that depends on factors such as the timing and duration of exposure (ie, repeated exposure prior to cocaine self-administration versus acute pretreatment prior to cocaine self-administration) and the reinforcement schedule used. Repeated THC exposure prior to access to cocaine self-administration can either attenuate, increase, or have no effect on the reinforcing effects of cocaine, although some of these mixed findings may be attributed to methodological differences, such as the length of THC exposure, the age of the rodents, and the parameters of self-administration. Acutely, THC also appears to exert inconsistent effects on cocaine reinforcement, such that it reduced cocaine CPP in rodents but increased cocaine choice in a nonhuman primate model. CBD, on the other hand, appears to either have no effect or have favorable effects on cocaine reinforcement, but decreases in cocaine-induced CPP.

2. Methamphetamine

Like studies examining cocaine and cannabis co-use, most preclinical behavioral studies investigating methamphetamine and cannabis co-use examined CBD or THC in isolation. One study found that after methamphetamine-induced CPP was established in rats, CBD administration (intraperitoneal) 1 hour prior to methamphetamine administration reduced methamphetamine-induced CPP in a dose-dependent manner.²⁸⁵ Furthermore, CBD administration facilitated CPP extinction following discontinuation of methamphetamine injections.²⁸⁶ Another study found that intracerebral infusions of CBD 1 hour prior to methamphetamine administration suppressed methamphetamine-induced reinstatement after CPP extinction; this suppression of reinstatement occurred in both healthy rats and those that were rapid eye movement sleep-deprived.²⁸⁷ These findings have been replicated by others, who found that CBD facilitated the extinction of methamphetamine-induced CPP and prevented reinstatement after extinction.288, 289, 290, 291 In addition, recent studies have shown that administration of CBD prevented the acquisition of methamphetamine-induced CPP.292, 293, 294 Another study examining how acute intraperitoneal CBD influenced methamphetamine self-administration at one dose (0.1 mg/kg) in rats found that CBD dose-dependently decreased the number of methamphetamine infusions on an FR schedule and reduced methamphetamine-primed reinstatement after extinction. However, CBD was not effective in reducing methamphetamine infusions on a PR schedule of reinforcement.²⁹⁵

While the above studies examined CBD in isolation, one study examined the efficacy of an oral 1:1 ratio of CBD to THC in attenuating methamphetamine-induced CPP in mice.²⁹⁶ On testing days after methamphetamine-induced CPP was established, the combination of CBD and THC decreased CPP.²⁹⁶ Another study examining THC alone reported that administering THC to rats after methamphetamine extinction sessions suppressed the reinstatement of methamphetamine-seeking behavior following a methamphetamine priming injection. Additionally, THC alone did not result in reinstatement at any of the doses tested. However, when THC was co-administered with a subthreshold methamphetamine dose, one that did not induce reinstatement on its own, it significantly increased responding.²⁹⁷ This may suggest that while THC alone does not influence methamphetamine-induced reinstatement, its co-administration produces a biphasic modulation of methamphetamine’s reinstating effects: enhancing reinstatement at subthreshold methamphetamine doses, but attenuating reinstatement at higher doses.

Although not directly assessing THC or CBD, several studies have investigated how cannabinoid receptor agonists and antagonists influence methamphetamine reinforcement. In one study, a cannabinoid receptor type 1 (CB1 receptor) antagonist significantly and dose-dependently reduced methamphetamine self-administration on an FR schedule.²⁹⁸ In contrast, pretreatment with the endogenous cannabinoid ligand anandamide had no significant effect, though a non-significant, dose-dependent trend toward increased methamphetamine self-administration was observed.²⁹⁸ The findings that cannabinoid receptor antagonists may have efficacy in reducing methamphetamine reinforcement have been replicated several times.298, 299, 300 Given that CBD may act as a functional antagonist of the endocannabinoid system, this phytocannabinoid, along with other CB1-targeting compounds, warrants further investigation as a potential therapeutic for MUD.²⁹¹ This line of inquiry is especially promising, given that multiple studies have also reported neuroprotective effects of CBD, THC, and synthetic CB1 agonists against methamphetamine-induced neurotoxicity.301, 302, 303

In sum, preclinical evidence suggests that cannabinoid-based compounds, particularly CBD and CB1 receptor antagonists, hold promise in reducing methamphetamine-related behaviors, including reinforcement, extinction, and reinstatement. While CBD has shown consistent efficacy in attenuating methamphetamine-induced CPP and in reducing self-administration under certain conditions, its effects do not persist under more demanding schedules of reinforcement. THC, in contrast, appears to exert more complex effects, with the potential to both suppress and enhance methamphetamine reinforcement, depending on whether the drug is co-administered with methamphetamine. The involvement of the endocannabinoid system is more broadly underscored by studies showing that CB1 antagonists reduce methamphetamine self-administration, while agonists enhance it.

3. Summary

Preclinical evidence shows that cannabinoids exert varying effects on stimulant reinforcement. For cocaine, THC exposure can increase, decrease, or have no effect on reinforcement, and acute THC reduces cocaine-induced CPP in rodents but increases cocaine choice in nonhuman primates. CBD generally has neutral or attenuating effects on cocaine reinforcement. CBD also consistently decreases methamphetamine-induced CPP, facilitates extinction, and reduces reinstatement. While THC does not influence high-dose methamphetamine-induced reinstatement, it does increase methamphetamine-induced reinstatement when coadministered with a low dose of methamphetamine that is ineffective on its own. Overall, THC has mixed effects on cocaine and methamphetamine reinforcement, while CBD reliably decreases stimulant-related behaviors.

E. Stimulants and opioids

1. Cocaine

As noted previously, stimulants such as cocaine are playing an increasingly dangerous role in the ongoing drug epidemic, particularly when used in combination with opioids.³ As a result, numerous studies have examined how various opioids, most notably heroin, influence cocaine reinforcement in animal models.⁴⁶ Studies examining the discriminative stimulus effects of cocaine have found that pretreatment with a low dose of heroin in monkeys resulted in leftward shifts of the cocaine dose-response curves. However, these effects did not reach statistical significance.^304,305 Similar results have been observed in rodents, where combinations of heroin and cocaine produced discriminative stimulus effects comparable to those of heroin or cocaine alone.³⁰⁶ Although there is limited evidence that heroin enhances the discriminative stimulus effects of cocaine, a study in rats demonstrated that pretreatment with morphine significantly potentiated these effects. This suggests that, under certain conditions, μ-opioid receptor agonists may enhance the interoceptive effects of cocaine.³⁰⁷

Early studies assessing speedball (heroin and cocaine combinations) self-administration in monkeys demonstrated that dose-response curves for cocaine and heroin combinations on a second-order schedule of reinforcement were not significantly different from those of cocaine and heroin alone.³⁰⁴ These findings have been replicated in studies involving rodents, such that combinations of heroin and cocaine produced reinforcing effects similar to either heroin or cocaine alone.³⁰⁶ Additional research using a PR schedule of reinforcement in monkeys showed that heroin increased the potency of cocaine to serve as a reinforcer without altering its reinforcing strength.³⁰⁸ While a rodent study replicated this pattern, another found that heroin enhanced both the potency and reinforcing strength of cocaine on a PR schedule of reinforcement.^309,310 These discrepancies are likely due to methodological differences, such as variations in dosing or reinforcement schedules.

Furthermore, studies using behavioral economic approaches have shown that when monkeys were allowed to smoke heroin, cocaine, or a combination of both, demand was greater for the combination compared with heroin alone, but did not significantly differ from demand for cocaine alone.³¹¹ Similar results have been generated in studies examining the oral administration of cocaine and the μ-opioid receptor agonist methadone, such that monkeys who had concurrent access to cocaine alone or cocaine and methadone robustly preferred the drug combination over cocaine alone.³¹² Taken together, these studies suggest that opioid-cocaine combinations can alter the discriminative stimulus and reinforcing effects of cocaine in nuanced and context-dependent ways.

An important consideration is whether opioid dependence or withdrawal modulates the reinforcing effects of cocaine. One study in monkeys found that chronic (67–123 days) daily administration of morphine 21 hours prior to cocaine self-administration on an FR schedule did not influence the potency of cocaine.³¹³ Similarly, a study of rats showed that methadone maintenance did not alter cocaine self-administration on an FR schedule of reinforcement.³¹⁴ The premise that opioid dependence does not modulate cocaine self-administration has also been recently replicated in rodents responding under a PR schedule of reinforcement during morphine dependence.³¹⁵ Additionally, a study evaluating the effects of methadone maintenance using osmotic minipumps in a rodent model of cocaine-induced reinstatement found that methadone maintenance blocked cocaine-induced reinstatement.³¹⁶

Studies examining the role of opioid withdrawal and cocaine co-use have been mixed. One study of monkeys using behavioral economics found that self-administration of cocaine was higher during periods of morphine withdrawal.³¹⁷ In rodents, the acquisition of cocaine reinforcement after one dose was not influenced by the discontinuation of chronic morphine treatment; cocaine self-administration after one dose was increased on a PR schedule 5 or more days after the onset of morphine withdrawal.⁵⁰ Another study also found that oxycodone withdrawal increased cocaine self-administration in rats responding under an FR schedule.³¹⁸ However, another study found that in morphine-withdrawn rats, cocaine self-administration under an FR schedule was not altered.³¹⁵ The latter findings were extended to morphine-dependent rodents responding under a PR schedule of reinforcement.³¹⁵

2. Methamphetamine

Several studies have examined the behavioral interactions between various opioids and methamphetamine. Most recent studies have focused on the co-use of fentanyl and methamphetamine. One study examining sex differences in the combined use of methamphetamine and fentanyl in rats found that when fentanyl was self-administered alone on an FR schedule of reinforcement, the session duration to earn all the reinforcers increased with increasing fentanyl doses.³¹⁹ However, when one dose of methamphetamine was added to these doses of fentanyl, this interaction was no longer present. In males, the latency to first infusion was also longer when fentanyl was available alone compared with fentanyl and methamphetamine combined. These findings suggest that methamphetamine modulated some of the rate-decreasing effects of fentanyl and facilitated faster responding in male rats.³¹⁹ A recent study also found that after extinction training of methamphetamine-induced CPP, low doses of methamphetamine that were ineffective on their own successfully reinstated drug-seeking behavior when combined with fentanyl.³²⁰

Another study found that when rats were given concurrent access to methamphetamine (0.1 mg/kg) and fentanyl (0.0032 mg/kg), their overall group-level response was similar to the levers associated with each drug.³²¹ However, more detailed analyses revealed 3 distinct patterns of behavior: some rats showed an almost exclusive preference for methamphetamine or fentanyl; some rats alternated their preference across days; and a third group consistently responded to both drugs within each session. Moreover, when the cost of one drug was increased (by decreasing its dose), the rats generally shifted their behavior toward the alternative drug, suggesting that methamphetamine and fentanyl could function as imperfect substitutes.³²¹ Although the reasons for these 3 patterns of drug co-use remain unclear, the findings suggest meaningful individual variability in how these drugs are co-used. This variability will likely be important to consider when developing effective interventions. Another study by this same group found that when rats were morphine-dependent or in morphine withdrawal, self-administration of methamphetamine was unchanged under a PR schedule of reinforcement.³¹⁵ A study examining methamphetamine and fentanyl co-use on a drug-versus-food choice procedure found that in non–opioid-dependent rats, the potency of the drug combination was significantly higher than either drug alone. When opioid dependence was produced, opioid withdrawal signs were significantly associated with increased fentanyl and methamphetamine choice for several days.³²²

Similar to fentanyl, morphine has also been shown to increase methamphetamine-induced CPP when administered in combination, compared with either methamphetamine or morphine alone.³²³ Interestingly, another study found that methamphetamine pretreatment increased the potency of morphine and methadone analgesia, as measured via a talk-flick assay in mice. These findings suggest that opioids may not only increase the rewarding effects of methamphetamine but may also potentiate the analgesic effects of opioids.³²⁴ Thus, methamphetamine and opioids may facilitate co-use by producing synergistic effects that amplify both reward and pain relief.

3. Summary

In summary, the effects of opioids on cocaine reinforcement appear mixed, likely due to the complex interactions between the specific opioid used, the timing and pattern of administration, and the behavioral paradigms employed. While some studies suggest that opioids can increase the discriminative stimulus effects of cocaine and increase the reinforcing potency and strength of cocaine, these effects are not consistently observed across studies. Interestingly, most studies support the view that chronic opioid administration does not alter the reinforcing potency or strength of cocaine. These findings suggest that maintenance therapies with opioid agonists, such as methadone, are unlikely to heighten sensitivity to cocaine use—an important consideration for individuals undergoing treatment for OUD. Additionally, the role that opioid withdrawal plays in cocaine reinforcement remains unclear. For methamphetamine co-use with opioids, research has demonstrated that methamphetamine and various opioids likely interact to produce effects that increase reward and drug-seeking behavior, potentially amplifying the analgesic effects of opioids. Additionally, when animals were opioid dependent, opioid withdrawal signs were significantly associated with increased fentanyl and methamphetamine choice for several days.

Some of the findings suggesting that co-administered opioids can increase the reinforcing effects of cocaine are consistent with a positive reinforcement mechanism driving co-use. By contrast, other observations, such as increased cocaine intake or preference during opioid withdrawal, may reflect negative reinforcement, in which co-use alleviates aversive states associated with opioid withdrawal. Similarly, studies investigating methamphetamine and opioid co-use support the idea that both positive and negative reinforcement mechanisms may drive co-use of stimulants with opioids.

F. Main takeaways

Across preclinical models, the effects of stimulants are strongly shaped by the co-use of other drugs, though the direction and magnitude of these interactions vary (summarized in Table 3). Nicotine increases the potency, the discriminative stimulus effects, and the reinforcing strength of cocaine. However, the effects of nicotine on methamphetamine are more conditional and are influenced by developmental timing, sex, and prior exposure. Alcohol shows mixed effects with cocaine and methamphetamine and may potentiate or attenuate stimulant reinforcement, depending on dose, timing, and experimental procedure. Cannabinoids also exert mixed effects: THC can increase or decrease stimulant reinforcement, while CBD reliably attenuates cocaine and methamphetamine reinforcement, reinstatement, and CPP. Finally, opioids may potentiate cocaine and methamphetamine reinforcement, especially during opioid withdrawal. Overall, stimulant polysubstance interactions in preclinical models are highly complex and likely influenced by interacting pharmacological, behavioral, and contextual factors. Because many studies do not fully account for variables such as drug dose, timing, route of administration, and prior exposure, findings often appear contradictory across experiments.

Table 3.

Summary of preclinical findings on polysubstance interactions between cocaine or methamphetamine and commonly co-used substances

Co-Used Substance	Cocaine-Related Findings	Methamphetamine-Related Findings
Nicotine	Nicotine increased cocaine reinforcement, potentiated cocaine’s discriminative stimulus effects, and increased the CPP effects of cocaine.237, 238, 239, 240, 241, 242, 243 Nicotine interfered with the efficacy of a behavioral intervention.243	Nicotine increased methamphetamine reinforcement in several studies, but this is shaped by a variety of other variables.244, 245, 246, 247, 248, 249, 250, 251, 252, 253 Nicotine mitigated methamphetamine-induced cognitive deficits.254,255
Alcohol	Mixed effects based on experimental paradigm.257, 258, 259, 260, 261, 262, 263, 264, 265, 266, 267, 268, 269, 270, 271 Alcohol reduced the efficacy of a behavioral intervention.272	Mixed effects based on experimental paradigm, but methamphetamine may potentiate methamphetamine-induced neurotoxicity.275, 276, 277, 278, 279
Cannabis	Mixed effects of THC based on experimental paradigm.280,281,283,284 CBD has neutral or attenuating effects on cocaine.282,284	Mixed effects of THC based on experimental paradigm.296,297 CBD reliably decreased the effects of methamphetamine.285, 286, 287, 288, 289, 290, 291, 292, 293, 294, 295, 296
Opioids	Mixed effects based on experimental paradigm.304, 305, 306, 307, 308, 309, 310, 311, 312 Chronic opioid administration does not alter cocaine reinforcement.313, 314, 315, 316	Opioid co-use increased the effects of methamphetamine and potentiated the analgesic effects of opioids.319, 320, 321, 322, 323, 324

VI. Conclusions

The evolving stimulant use crisis has increasingly become characterized by complex patterns of polysubstance use. Despite its profound consequences for public health, polysubstance use has historically received limited research attention, leaving major gaps in understanding and treating StUDs. This review highlights that stimulants are rarely used in isolation. Instead, they are commonly combined with substances such as alcohol, nicotine, cannabis, and opioids, and these combinations can amplify rewarding, subjective, and reinforcing effects and have been associated with health risks due to pharmacodynamic and PK interactions. In fact, mechanistic evidence shows that co-use can enhance dopaminergic signaling, prolong stimulant bioavailability, and alter absorption and metabolism in ways that increase the likelihood of combined use. These interactions may increase desired drug effects, like euphoria, while blunting unwanted side effects such as sedation or anxiety. Importantly, stimulant co-use likely arises from both positive and negative reinforcement processes. While many studies demonstrate how co-use potentiates stimulant reward and reinforcement, others suggest that continued or sequential use may serve to mitigate dysphoria, anxiety, or physiological withdrawal. Recognizing both motivational pathways helps reconcile conflicting findings across studies and underscores that stimulant polysubstance use is not a unitary phenomenon but rather an interplay between behaviors maintained by positive and negative reinforcement.

When considered together, the literature reviewed here suggests that the translational value of polysubstance research will be maximized when studies align with how co-use is defined and modeled across epidemiological, clinical, and preclinical domains. Although many studies already measure aspects of timing, dose, route, or motivation, these features are often examined in isolation or emphasized differently across epidemiological, preclinical, and clinical research. The studies discussed highlight a common set of parameters that are most likely to shape drug interactions and outcomes, including the temporal pattern of use (separate, simultaneous, or sequential), the functional role of the co-used substance (positive or negative reinforcement), the route of administration (and the resulting PK), the doses used, and the stage of stimulant use (acquisition, maintenance, treatment, and relapse) (Fig. 2). Importantly, when feasible, these co-use dimensions should be considered alongside other factors that shape risk and treatment responses, including, but not limited to, sex, the social environment, and developmental stage. This framing enables researchers to translate epidemiological descriptions of polysubstance use into specific experimental variables (timing, dose, route, and motivation) to test how these factors impact the acquisition, maintenance, treatment, and relapse risk of stimulant use.

Fig. 2.

Schematic overview of a common set of dimensions that are most likely to influence drug interactions and outcomes in stimulant polysubstance use. When feasible, these co-use dimensions should be considered alongside other factors that shape risk and treatment response, including sex, social environment, and developmental stage.

A. Major gaps to be addressed

1. Epidemiological gaps

Epidemiological research has been essential in establishing that polysubstance use is the norm rather than the exception among people who use stimulants and in identifying which substance combinations are associated with elevated health risks, overdose, and psychiatric comorbidities. However, when considered alongside preclinical and clinical work, a key limitation becomes apparent. Most epidemiological studies are largely descriptive and are currently well-suited to identifying which drugs are commonly co-used and what the risks of polysubstance use are. However, they rarely determine the temporal pattern of co-use, the motivations for co-use, or the route and doses used across the stimulant use cycle. For instance, current surveillance systems rarely capture dimension of co-use in sufficient detail to inform clinical or preclinical models, and consequently, researchers lack a clear picture of whether stimulant co-use typically occurs sequentially (eg, alcohol before cocaine), simultaneously (eg, combined smoking of methamphetamine and heroin), or temporally separate (using methamphetamine hours after opioid administration to reduce withdrawal). Importantly, this is an active area of investigation, and several research groups are already developing methods to quantify key variables related to polysubstance use, such as co-use patterns over time.^94,325 This work provides a strong foundation that should be expanded and more consistently integrated into study designs. Moreover, large-scale, longitudinal studies are needed to systematically document co-use motivations, quantities/doses, and routes, as well as how these factors evolve across stages of initial use, maintenance, treatment, and relapse. These data are essential for grounding preclinical and clinical designs in real-world use patterns and for identifying which combinations pose the greatest risks for continued use despite negative consequences and treatment failures (Fig. 3).

Fig. 3.

Schematic overview of key gaps in the current literature and the need for stronger integration of preclinical, clinical, and epidemiological research. Enhanced cross-talk across these domains will facilitate complementary findings, close critical knowledge gaps, and accelerate the development of effective treatments for stimulant-related polysubstance use.

2. Human subject gaps

In clinical trials, stimulant polysubstance use remains underrepresented in studies examining interventions for StUDs. Even though people who use stimulants rarely take only one drug at a time, most treatment studies still enroll or analyze participants as if they do, leaving major uncertainties about how co-use alters treatment access, retention, and efficacy. For cocaine and methamphetamine use, smoking cigarettes and cannabis is ubiquitous, yet nicotine or cannabis use is rarely quantified or controlled for in clinical trials. This leaves a major blind spot and makes it harder to interpret variability in treatment outcomes. Gaps are also evident for stimulant and opioid interactions. Clinical trials of CUD or MUD rarely distinguish between participants who used opioids intermittently, were dependent but untreated, or were chronically maintained on an opioid agonist as a therapeutic for OUD. In short, clinical trials often collapse findings across subjects despite heterogeneity in co-use temporal patterns, motivations for co-use, route of administration, and doses used.

As noted previously, many of these gaps may be because recruiting participants who use stimulants only as a control group is impractical since polysubstance use is extremely prevalent. Rather than avoiding the study of polysubstance use altogether, clinical trials could be designed to pre-specify subgroups based on patterns and severity of co-use. For example, studies could compare individuals who co-use cocaine and alcohol but do not engage in binge drinking while using cocaine with those who do. When these subgroup comparisons are defined a priori and incorporated into the statistical analysis plan, they allow for the evaluation of whether specific co-use patterns moderate treatment response or relapse outcomes. Although such approaches require adequate sample sizes within subgroups, incorporating them into clinical trials would improve the interpretation of treatment efficacy data for individuals StUDs (Fig. 3). Encouragingly, recent work on large treatment-seeking cohorts has begun to empirically derive polysubstance use profiles (for example, moderate polysubstance use versus high past-month polysubstance use) and link them to clinically meaningful outcomes, such as overdose and mental health comorbidities.³²⁶ This provides a foundation for future studies to test whether such polysubstance use profiles moderate treatment outcomes.

Human laboratory studies often investigate many key parameters of polysubstance use, such as the sequence and timing of co-use, controlled dosing, subjective reports of motivation and drug effects during co-use, and routes of administration that reflect naturalistic patterns (Fig. 2). A major constraint, however, is that ethical considerations generally limit these studies to individuals who are already using stimulants, often under maintenance conditions. As a result, questions about initiation, relapse risk, and outcomes in treatment-seeking populations are often better addressed using preclinical models and epidemiological approaches.

3. Preclinical gaps

Preclinical studies have advanced our understanding of stimulant co-use mechanisms, and notably, several recent reviews discussing polysubstance use have stressed the need to incorporate drug co-use into animal models of substance use disorder.^221,236,327 However, when viewed through a translational lens, major gaps remain in how preclinical models operationalize key dimensions of polysubstance use (Fig. 2). First, most animal models still rely on simplified 2-drug paradigms (eg, cocaine and alcohol), which fail to capture the multidrug patterns common in human polysubstance use. While developing more complex models is an important long-term goal, expanding prematurely risks exaggerating existing methodological limitations. In fact, while several studies have established full dose-response curves to investigate polysubstance use, most current studies test only single-dose combinations, making it difficult to determine whether observed effects reflect true pharmacological interactions or simple additive changes in reinforcement. Strengthening the preclinical foundation through systematic full dose-response curve analyses, incorporating social and biological moderators, and implementing diverse reinforcement schedules beyond simple FR contingencies is essential before progressing to multidrug paradigms that more accurately mirror human co-use. Another major limitation involves determining the directionality of interactions between co-used drugs. For example, when 2 drugs with reinforcing effects are co-administered, it is often unclear whether one enhances the effects of the other or whether the observed effect is simply additive. Few studies systematically manipulate both drugs across a full dose-response range, making it challenging to parse these mechanisms. Clarifying these relationships will be an important future direction.

Closely related is the limited consideration of temporal patterns in preclinical models. Across studies, co-use is operationalized in highly variable ways. For example, some paradigms administer drugs in fully temporally separate sessions, others model simultaneous co-administration, and others use sequential dosing with varying inter-drug intervals. Importantly, temporal patterns used in preclinical studies do not always align with those suggested by epidemiological studies. In many cases, however, this likely reflects the fact that real-world sequences of co-use are not yet well characterized. This gap underscores the need for continued cross-talk between research domains, particularly for more detailed epidemiological data on timing, so that preclinical models can better align temporal patterns of co-use with what is most often observed across stages of stimulant use, including acquisition, maintenance, treatment, and relapse.

Within existing models, several compound-specific limitations persist. For cocaine and alcohol, mixed outcomes ranging from ethanol increasing, decreasing, or having no effect on cocaine reinforcement likely reflect ethanol’s broad pharmacological profile, variability in blood concentration, and the in vivo formation of cocaethylene.328, 329, 330, 331 Rigorous assessment of these parameters will be critical for clarifying the direction and mechanisms of interactions. Another overarching limitation concerns the underrepresentation of negative reinforcement mechanisms across co-use paradigms. This omission likely stems from the difficulty of modeling stimulant withdrawal-related symptoms such as anhedonia and dysphoria, which are central to StUDs but challenging to capture preclinically. Nonetheless, advancing procedures capable of assessing negative reinforcement are necessary for a comprehensive understanding of polysubstance use.

For cannabis and stimulant interactions, ecological validity remains limited due to a mismatch between preclinical and human drug administration. In humans, cannabis is typically smoked and contains more than 550 chemical compounds and over 100 phytocannabinoids, but preclinical studies predominantly rely on injectable THC or CBD formulations.³³² While preclinical studies have begun to incorporate vaporized or smoked delivery methods of full-spectrum cannabinoid extracts,^333,334 these methods have not yet been incorporated into models of polysubstance use. The same principle applies to stimulant and nicotine models, as most preclinical work examining polysubstance use employs intravenous nicotine administration, which differs from the PK of smoked or vaped nicotine. Although detailed consideration is beyond the scope of this review, future work should also address how other widely used stimulants, such as caffeine and 3,4-methylenedioxymethamphetamine, interact with cocaine and methamphetamine in polysubstance use contexts.

Beyond these substance-specific issues, preclinical studies rarely evaluate behavioral or pharmacological interventions under polysubstance use conditions during the treatment stage of stimulant use. Instead, most preclinical polysubstance research focuses on how co-use influences the acquisition and maintenance of drug use and reinstatement after abstinence. This imbalance represents a major translational gap because preclinical models are the critical first step in identifying candidate medications for StUDs before they advance to human testing. Without evaluating potential treatments in paradigms that capture the interactive effects of multiple drugs, it remains unclear whether compounds that appear effective under single-drug conditions will maintain efficacy under the polysubstance use patterns that dominate clinical populations. Furthermore, expanding beyond 2-drug designs toward multidrug paradigms (eg, cocaine + alcohol + nicotine) will improve clinical relevance but must be coupled with greater methodological rigor and reproducibility. Together, these refinements will strengthen the translational bridge between preclinical and clinical research, enabling more accurate modeling of human stimulant polysubstance use and its treatment (Fig. 3).

B. Recommendations for future research

Beyond identifying gaps across individual research domains, a key challenge is integrating mechanistic, epidemiological, preclinical, and clinical findings into a coherent framework that can inform study design. Epidemiological studies are well-positioned to identify which substances are commonly co-used and which combinations are associated with elevated risk in the stimulant use cycle. However, they often lack detailed information on sequencing, dose, route, and motivation. Conversely, preclinical and human laboratory studies can manipulate these variables but are often designed without a clear grounding in real-world patterns of use. Clinical trials, in turn, document heterogeneity in treatment responses among people with StUDs, but these differences are rarely linked to specific features of polysubstance use.

Future research should explicitly bridge the epidemiological, preclinical, and clinical domains by considering a shared set of co-use dimensions that shape drug interactions and outcomes (Fig. 2). When feasible, studies should report or examine the pattern of co-use, the motivation for co-use, the route of administration, the doses used, and the stage of stimulant use being modeled (acquisition, maintenance, treatment, or relapse). Importantly, when incorporated into the study design, these variables should be based on real-world patterns. Accordingly, epidemiological studies should assess when and how substances are co-used, why they are combined, typical doses and ratios (for example, grams of cocaine alongside the number of standard drinks), and whether these parameters differ across initiation, maintenance, treatment-seeking, and relapse. These same dimensions can then be directly translated into human laboratory and preclinical models to test causal hypotheses by systematically manipulating each parameter. In parallel, clinical studies can evaluate whether these co-use features moderate craving, reinforcement, relapse risk, and treatment response.

Together, these coordinated approaches would move the field beyond isolated lines of evidence across preclinical, clinical, and epidemiological studies toward an integrated framework that explains how stimulant co-use emerges, is sustained, and can be targeted through interventions. Closing these gaps will require greater cross-talk between preclinical, clinical, and epidemiological researchers (Fig. 3). Funding agencies, journals, and professional organizations can accelerate progress by incentivizing interdisciplinary teams and encouraging the use of shared reporting frameworks for co-use parameters, since no single discipline can address the stimulant polysubstance use crisis in isolation.

Conflict of interest

No author has an actual or perceived conflict of interest with the contents of this article.

Associate Editor: Matthew Banks