Cannabis: concerns, origins and threats to the validity of abuse/misuse potential

Cannabis abuse is a term describing the continued pathological use of cannabis. In DSM-5, the American Psychiatric Association defines cannabis use disorder (CUD) in terms of discrete patterns classified under impaired control, social impairment, risky behaviour or physiological adaptation, all with varying degrees of intensity. ^{Reference Patel and Marwaha7}

Cannabis use disorder is said to affect about 10% of the 193 million cannabis users globally, and it encompasses the possibility that people may suffer adverse consequences because of cannabis use, i.e. beginning with, or leading to, misuse/abuse, without necessarily becoming addicted. Importantly and ironically, the struggle to establish criteria for a continuum of problematic use, abuse or misuse, dependence or addiction seems to contribute more to diagnostic and classificatory confusion rather than clarity. ^{Reference Connor, Stjepanovic, Le Foll, Hoch, Budney and Hall8}

Cooper and Abrams ^{Reference Cooper and Abrams9} have observed that there is relatively little quality research regarding conditions under which cannabis abuse and dependence are more likely to emerge. Schlossarek et al ^{Reference Schlossarek, Kempkensteffen, Reimer and Verthein10} undertook a literature search of 10 568 studies on the subject. Twenty-six of these finally met the inclusion criteria, and the group’s conclusions included an overarching statement about how striking it was that some studies used terms such as problematic use, abuse and dependence while lacking a precise operationalisation or using the term ‘dependence’ even if the required number of symptoms was not fulfilled. Often, important information on patterns of use, method of use or administration, quantity and quality (type, potency, purity and formulation) was absent, as well as details of context, motives and subjective effects. Importantly, there was little distinction between recreational use and recreational products versus medical use and medical cannabinoids (such as Marinol (dronabinol), the synthetic Δ ⁹-tetrahydrocannabinol (Δ ⁹-THC) approved in 1985 by the FDA for chemotherapy-associated nausea and anorexia.

Analysis of the strongest studies seemed to yield factors that were associated with an impact on the transition from cannabis use to dependence. Perhaps not surprising was that a wide range of psychiatric disorders was associated with an elevated risk of becoming dependent. The development of a dependence syndrome was also linked with an aetiology in which social, biological and intra-individual factors interact in a complex fashion, leading to dependence. Nevertheless, the putative link between cannabis dependence and predisposing factors was not readily resolved by most studies, due to definitional and methodological weaknesses regarding dependence criteria.

Thus, it is suggested that CUD may more accurately be a proxy for an overarching psychiatric diagnosis that creates psychosocial vulnerabilities for what has been traditionally termed substance abuse. Another important challenge to a broad characterisation of ‘abuse liability’ that lumps every cannabis user into a potential abuse/misuse category lies in genomic variability. Johnson et al ^{Reference Johnson, Demontis, Thorgeirsson, Walters, Polimanti and Hatoum11} conducted a large-scale, genome-wide association study meta-analysis of cannabis use disorder and found associations with two genome-wide significant loci: a novel chromosome 7 locus (FOXP2, lead single-nucleotide polymorphism (SNP) rs7783012) and the previously identified chromosome 8 locus (near CHRNA2 and EPHX2, lead SNP rs4732724). Cannabis use disorder was also positively genetically correlated with other psychopathology, including attention-deficit hyperactivity disorder, major depression and schizophrenia, consistent with the findings of Schlossarek et al. ^{Reference Schlossarek, Kempkensteffen, Reimer and Verthein10}

Gender differences in response to cannabis exposure suggest yet another source of variability. Cooper and Haney ^{Reference Cooper and Haney12} found that, among cannabis smokers, men exhibit greater cannabis-induced analgesia compared with women. Such a gender-dependent difference appears to be independent of the cannabis-elicited psychoactive effects associated with promotion of abuse liability. Thus, gender-dependent differences in cannabis’s analgesic effects appear, at least, to be an important research pathway that justifies further investigation.

Variability in dose and route of administration

The challenge of dose titration in the medical use of cannabis is complicated by the high degree of variability observed in individual responses to ingested Δ ⁹-THC. Clinical experience with dronabinol, a synthetic isoform of Δ ⁹-THC, suggests that, for some individuals, 2.5 mg is sufficient to produce apparent effects. In contrast, higher doses are necessary for others – in some cases exceeding 50 mg (reviewed in Grotenhermen ^{Reference Grotenhermen13} ). In light of this variability, probably influenced by gender, age and genomics, calculation of pharmacologic potency between a given quantity of Δ ⁹-THC contained in smoked cannabis and a mass of Δ ⁹-THC contained in an edible product, for example, is extremely difficult.

First-pass metabolism is also a potentially significant source of dose variability; liver enzymes hydroxylate Δ ⁹-THC to form 11-hydroxytetrahydrocannabinol (11-OH-THC), a more potent psychoactive metabolite that readily crosses the blood–brain barrier (BBB). ^{Reference Mura, Kintz, Dumestre, Raul and Hauet14} 11-OH-THC is more potent than Δ ⁹-THC, ^{Reference Hollister, Gillespie, Ohlsson, Lindgren, Wahlen and Agurell15,Reference Huestis, Henningfield and Cone16} and its serum levels are higher when Δ ⁹-THC is ingested orally than when it is inhaled. ^{Reference Grotenhermen13} In addition to inhaled (smoked, vaporised or aerosolised) and oral administration, administration of medical cannabinoids via other modes is being explored. Oral-mucosal, sublingual, transdermal, transrectal and transvaginal formulations are under preclinical investigation. Each mode of drug delivery has its own spectrum of complexity and, as yet, incomplete characterization of its absorption/distribution/metabolism/excretion, and it will be interesting to evaluate these data as they emerge.

Withdrawal

Withdrawal is an essential component of classical addiction theory and is a major criterion used to determine whether an agent is addictive or subject to recurrent abuse or misuse liability; it is a vital manifestation of dependence. ^{Reference Piper17} The physical dependence on substances of abuse/misuse is phenomenologically likely to involve repeated cycles of withdrawal accompanied by increased motivation to self-administer that substance. This phenomenon is generally not seen with medical cannabis, and is unusual in the realm of recreational cannabis use. There is longstanding evidence, however, of a withdrawal syndrome reliably following abrupt discontinuation of chronic heavy cannabis use. ^{Reference Budney, Hughes, Moore and Vandrey18}

Withdrawal symptoms in patients with neuropathic pain treated with the cannabinoid dronabinol or Δ ⁹-tetrahydrocannabinol, a synthetic form of tetrahydrocannabinol (THC), included only mild and transient sleep disturbance, anxiety and increase in neuropathic pain. ^{Reference Schimrigk, Marziniak, Neubauer, Kugler, Werner and Abramov-Sommariva19}

The Schimrigk trial was designed to explore the positive benefit/risk ratio of dronabinol among 240 multiple sclerosis patients with central neuropathic pain. The trial protocol was a 16-week, placebo-controlled Phase III study followed by a 32-week, open-label period. One hundred patients received therapy for up to 119 weeks. The primary end-point was a change in pain intensity throughout the treatment period. Safety evaluation was based on sequelae of dependency, and abuse. No signs of drug abuse, and only one possible case of dependency, were reported. Consistent with decades of clinical experience with Marinol or dronabinol, the trial suggests safety over a significant duration.

For most patients, no withdrawal reactions were reported following cessation of study medication: withdrawal symptoms were reported for six patients following the open-label period and for four patients after the extended follow-up. Importantly, diagnostic criteria for drug dependency and abuse were serially assessed by investigators during the long-term follow-up period. Mild signs of drug dependency were reported for one patient.

In another older study, for clinical trials of Sativex (a THC + cannabidiole (CBD) medicine), intoxication scores were low and euphoria was reported by only 2.2% of subjects. Tolerance did not occur, abrupt withdrawal failed to show a stereotypic withdrawal syndrome and no cases of abuse were reported in follow-up. A formal abuse liability study of Sativex in long-term cannabis smokers suggested some abuse potential in comparison with a placebo. ^{Reference Robson20}

Synthetic cannabinoids

It has been emphasised that much of our albeit flawed understanding of cannabinoid tolerance, dependence and withdrawal has been based on experience with Δ ⁹-THC, a relatively weak partial agonist at CB1 and CB2 receptors. ^{Reference Tai and Fantegrossi21} However, the synthetic cannabinoids (SCBs) commonly found in commerce and on the ‘black market’, such as K2 and Spice, are typically full cannabinoid receptor agonists. The authors point out that a low-efficacy cannabinoid like Δ9-THC will have a less pronounced maximal effect than higher-efficacy cannabinoids, such as the SCBs, and this difference in maximal effects cannot be overcome simply by increasing the dose of Δ ⁹-THC. In other words, no amount of Δ ⁹-THC can stimulate cannabinoid receptors to the same degree as the SCBs currently emerging as a principal subclass of cannabinoid drugs of abuse. This situation has left investigators working with efficacy SCBs in the difficult position of having to tease out whether the effects are related to the significant degree of cannabinoid receptor stimulation produced by these compounds, or whether interactions with other non-cannabinoid receptors or even non-receptor systems produce them. Thus, there is greater confusion in the increasingly dense ‘fog of war’ in understanding cannabinoid use and possible abuse/misuse.

Cannabis as a ‘gateway’ drug

The question of cannabis as a gateway drug is integral to the discussion of the linkage of this substance to other established drugs of abuse/misuse. Any gateway effect is more likely to be driven by social, psychological and biological vulnerabilities, not by the pharmacological properties of cannabis itself. There is no proven causal link between cannabis use for medical indications and progression to more potent or dangerous agents, ^{Reference Noel and Wang22} and medical cannabis often substitutes for opioids and other medications. ^{Reference Boehnke, Scott, Martel, Smith, Bergmans and Kruger23,Reference Reiman, Welty and Solomon24}

A report prepared by the Federal Research Division, Library of Congress, under an Interagency Agreement with the Office of the Director, National Institute of Justice, Office of Justice Programs, U.S. Department of Justice in November 2018 (https://www.ojp.gov/pdffiles1/nij/252950.pdf), stated:

‘The existing statistical research and analysis show mixed results and do not clearly demonstrate scientific support for cannabis use, leading to harder illicit drug use. As a result, FRD has determined that no causal link between cannabis use and the use of other illegal drugs can be claimed at this time.’

The report goes on to cite deficiencies or weaknesses in the reviewed research. It is noted that many studies of cannabis use are based on self-reported data from longitudinal or retrospective studies that suffer from bias; it seems axiomatic at this point that subjects often cannot accurately recall the specifics requested, tend to provide responses that they believe investigators want to hear or even portray themselves in a favourable light.

Other studies may collect biased data by sampling from heroin users, street youth and other at-risk populations. As such, the results are not readily generalisable. The limitations of animal models are also noted, along with the constraints of translating findings to human behaviour.

Finally, it is noted that 2018 statistics from the U.S. Department of Health and Human Services Substance Abuse and Mental Health Services Administration show that only 0.3% of those Americans who have used cannabis have abused heroin, 0.2% have abused cocaine and 0.1% have abused methamphetamines.

It also should be noted that the findings from another frequently cited study are consistent with those of the Federal Research Division report: it was found that the initiation of cannabinoid-based medicines in patients with chronic pain was associated with 17-fold higher odds of ceasing prescription analgesics within 21 months. ^{Reference Vigil, Stith, Adams and Reeve25}

On the other hand, one frequently cited study ^{Reference Secades-Villa, Garcia-Rodríguez, Jin, Wang and Blanco26} reported that 44.7% of individuals with lifetime cannabis use progressed to other illicit drug use at some time in their life. The fact remained, however, that several antecedent variables, the most important being underlying psychiatric disorder, were found to predict progression from cannabis use to other illicit drugs use.

Dosing mechanisms

Dose-response patterns in cannabinoids appear to be hormetic or biphasic, i.e. at lower doses there appear to be salutary effects whereas higher dose ranges are associated with undesirable and potentially hazardous outcomes. The mechanisms involved may also be either receptor- or non-receptor-based.

The consistent quantitative features of this biphasic dose response at the cellular level may be complicated by the effects of upstream and highly conserved gene clusters that control and direct the allocation of the metabolic programming of the entire organism. ^{Reference Calabrese27,Reference Calabrese and Mattson28}

Current evidence suggests that THC is anxiolytic at low doses and anxiogenic at higher doses, with each case demonstrating unique dose-response curves. Interestingly, CBD (possibly in conjunction with other cannabis plant constituents such as cannabinol, terpenes and flavonoids) may help to lower or antagonise the psychotropic effects of THC when used in particular combinations and proportions, an observation that underscores the priority for understanding the interaction between these two cannabinoids. ^{Reference Schoedel, Szeto, Setnik, Sellers, Levy-Cooperman and Mills29,Reference Stewart and Fong30}

Analyses of the overall safety of cannabinoid-based medicines are further confounded by variations in dosing and administration methods across products, study designs and indications. While there are standardised doses for CBD use in Lennox-Gastaut and Dravet syndromes, and in investigational studies, a commonly used approach to dosing is ‘start low, go slow, stay low’. ^{Reference MacCallum and Russo31,Reference Schuck, Pacanowski, Kim, Madabushi and Zineh32} While this ‘personalised’ or ‘precision’ medicine approach is helpful for patients as individuals, lack of standardisation, even within the same indication, represents a challenge for the overall analysis of safety, and further studies to establish appropriate standard doses or dose ranges for each indication are needed. Additionally, while many cannabinoid-based medicines are administered orally, inhalation – and, in particular, vaporisation, not smoking – may be deemed most appropriate. Table 2 summarises research on abuse potential that counters many of the conclusions and assumptions suggesting elevated abuse potential.

Table 2 Abuse characteristics of medical cannabis

RCT, randomised controlled trial.

Cannabis itself

There is little standardisation, consistency or harmonisation in commerce in regard to cannabis. This plant can provide over 500 different active chemical compounds that interact with numerous molecular targets across several receptor families, thereby modulating the production and metabolism of endocannabinoids, gamma-aminobutyric acid, glutamate and serotonin. Unlike 24 synthetics and alcohol, cannabis is a more complex drug(s). Consumption or inhalation of the botanical resin exposes the user to hundreds of compounds, including both cannabinoids (e.g. THC and cannabidiol) and non-cannabinoids (i.e. terpenes and flavonoids), many of which are bioactives. This is in direct contrast to isolated pharmaceutical derivatives (e.g. dronabinol and cannabidiol). The sheer complexity of the plant makes comparison difficult even between THC and CBD. ^{Reference Radwan, ElSohly, El-Alfy, Ahmed, Slade and Husni37}

Psychoactive effects are primarily derived from THC, which is a partial agonist at both cannabinoid CB1 and CB2 receptors. CB1 receptors are located throughout the central nervous system, lungs, liver and kidneys. Binding CB2 receptors modulates G-protein-coupled inhibition of cyclic adenosine monophosphate (cAMP), thereby influencing pain, mood, appetite and sexual activity. Interaction between exogenous cannabinoids and endocannabinoid tone is another fertile area for augmenting the understanding of mechanisms of action and the array of salutary and adverse effects.

It has been pointed out that the two most studied exogenous cannabinoids are THC and CBD. ^{Reference Campbell, Phillips and Manasco38} CBD does not appear to bind with significant affinity to CB1 or CB2 receptors but seems to possess independently modulated neuroprotective and anti-inflammatory effects. Several potential CBD actions have been proposed: inhibition of cyclooxygenase and lipoxygenase, inverse agonism at CB1/CB2 receptors and enhancement of anandamide (an endogenous THC analogue). It is proposed that CBD may be effective in epilepsy through modulation of the endocannabinoid system by halting the degradation of anandamide, which may have a role in inhibition of seizures. Additionally, CBD may play a role in regulation of T-type calcium channels and nuclear peroxisome proliferator-activated receptor-γ, both of which have been implicated in seizure activity.

Metabolites of cannabis

Hepatic metabolism can produce over 80 metabolites of Δ ⁹-THC, with the most common pathway involving allylic hydroxylation at the 11-position followed by oxidation to a carboxy derivative. Conjugation occurs with some metabolites, but it does not appear to be a major step. Bioavailability varies depending on the user’s smoking topography, such as number, duration, spacing of puffs, hold time and inhalation volume. It has long been known that THC remains in the body for extended periods due to its lipophilic properties, allowing it to accumulate and be released slowly from adipose tissue. ^{Reference Hollister, Gillespie, Ohlsson, Lindgren, Wahlen and Agurell15} It is tempting to speculate about delayed and unintended effects in those engaging in rapid weight loss.

It cannot be stressed too much that some of the differences and confounds between the properties of recreational versus medical cannabis are exemplified by the frequency of what appears to be sloppy comparison across various agents. Dronabinol, for example, contains only the synthetic version of Δ ⁹-THC whereas cannabis products may contain Δ ⁹-THC plus an array of cannabinoids and other compounds, including terpenes and cannaflavins. ^{Reference Russo39}

Variability in manufacture

Because of non-standard extraction and formulation, the amount of Δ ⁹-THC in edibles can vary across a single product type, and across batches processed at different times (brief shelf-life), making it challenging for users to estimate how much Δ ⁹-THC they consume. Apart from accessibility and the visual appeal of packaging, this fact may explain in part why some edible products are also associated with cannabis-related childhood poisoning.

Lower THC in the plasma may be the result of generally low bioavailability: although in a more potent form, the amount of THC that reaches the circulation following oral administration may be only be 6–10% of the amount contained in the product. ^{Reference Schwilke, Schwope, Karschner, Lowe, Darwin and Kelly40} The lack of product consistency, together with delayed intoxication (attributed to the time required for gastrointestinal absorption, possible food matrix and first-pass metabolism), may cause users of cannabis to consume higher than intended amounts of the drug. Edible products such as gummies are responsible for the majority of healthcare visits due to cannabis intoxication, which is often due to ignorance and inexperience in naïve users or their failure to appreciate the delayed effects. ^{Reference Vandrey, Raber, Raber, Douglass, Miller and Bonn-Miller41}

Variability in self-dosing

Dosage estimation for existing retail products is often inaccurate. ^{Reference Vandrey, Raber, Raber, Douglass, Miller and Bonn-Miller41} While state laws in the USA often require that total milligrams of Δ ⁹-THC and the number of servings be included on packages available for retail sale, doses are likely to vary widely. Individuals often tell a story about having eaten the suggested serving size initially, followed by a decision to consume the entire edible product after not feeling any effects. They also report that it was both easy and likely that they would consume the entire edible product in one sitting, just as they would a normal baked good, ^{Reference Russo39} suggesting at least the likelihood of, or vulnerability to, a persistent pattern of misuse.

Gottschling et al ^{Reference Gottschling, Ayonrinde, Bhaskar, Blockman, D’Agnone and Schecter42} provide additional emphasis that recreational cannabis use may be associated with uncertain product composition, with potentially high THC concentrations and unknown biologically active impurities. Again, dosing may be unknown and uncontrolled and usage may be accompanied by concomitant tobacco or other recreational drug use, each with its own set of independent and potential interactive effects.

Discussion