Cannabis, Psychedelics and New Psychoactive Substances
Drugs and Society
Learning objectives for this lesson:
- Describe the endocannabinoid system, including CB1 and CB2 receptors, anandamide, 2-AG and retrograde signalling, explain how THC and CBD differ in their actions and effects, and state how Canadian law classifies cannabis and its cannabinoids.
- Compare the cannabis products sold in Canada's legal market and the THC limits that apply to them.
- Summarize international and Canadian evidence that THC concentrations in cannabis have risen since the 1970s, state the limits of that evidence, and calculate the THC content of a product in milligrams and in standard THC units.
- Classify health effects of cannabis using the NASEM (2017) weight-of-evidence categories, and interpret the evidence on psychosis and cannabis use disorder.
- Explain the purpose of Canada's Lower-Risk Cannabis Use Guidelines and describe how the 2022 update changed the 2017 recommendations.
- Describe the receptor targets, acute effects and main risks of classic psychedelics, MDMA, ketamine and other dissociatives, and state their legal status in Canada as of 2026.
- Compare the dependence potential of these drugs with other drug classes using expert harm ratings, and identify the limits of such ratings.
- Define new psychoactive substances and describe how synthetic cannabinoids, synthetic cathinones and novel benzodiazepines act and cause harm.
- Explain how the EU Early Warning System, the UNODC Early Warning Advisory and Canadian monitoring detect, assess and report new psychoactive substances, and interpret Health Canada's data on substances identified for the first time, including their limits.
This course was developed by Dr. Kiffer G. Card.
Cannabis Pharmacology: The Endocannabinoid System, THC, CBD and Potency
Introduction
Canada legalized and regulated non-medical cannabis for adults when the Cannabis Act came into force on 17 October 2018, and regulations for edible cannabis, cannabis extracts and cannabis topicals followed on 17 October 2019 (Health Canada, 2019). Cannabis is therefore a legal product in Canada, sold in licensed stores with labels that state its THC content, and it is also a drug with well-studied effects on the brain. Lesson 3 introduced the shared pharmacology of receptors, agonists, potency and routes of administration, and Lessons 4 to 6 applied it to depressants, opioids and stimulants. The same tools apply to cannabis; Section 2 then weighs the evidence on its health effects, and Lesson 12 examines how legalization itself has worked.
Learning Objectives
- 1.1 Distinguish the four kinds of cannabinoid, and state how Canadian law classifies cannabis and its cannabinoids as of 2026.
- 1.2 Describe the parts of the endocannabinoid system and explain how endocannabinoids carry a retrograde signal across a synapse.
- 1.3 Distinguish THC from CBD by their actions and effects, including THC's action as a partial agonist and how route shapes the timing of effects.
- 2.1 Compare the main cannabis product forms sold in Canada's legal market and the THC limits that apply to them.
- 3.1 Summarize the evidence that THC concentrations have risen, and state the limits of that evidence.
- 3.2 Calculate the THC content of a product in milligrams and in standard THC units.
1.1 The Cannabis Plant and Its Cannabinoids
Health Canada describes cannabis as the plant Cannabis sativa, which contains hundreds of chemical substances, more than 100 of which are known as cannabinoids (Health Canada, 2023a). A cannabinoid is a compound that acts on, or is closely related to compounds that act on, the body's cannabinoid receptors. Health Canada names delta-9-tetrahydrocannabinol (THC) as the most researched cannabinoid and describes it as responsible for the high and intoxication that people associate with cannabis. Another cannabinoid, cannabidiol (CBD), does not produce a high or intoxication (Health Canada, 2023a).
The word "cannabinoid" covers several kinds of substance, which the box below separates.
Four kinds of cannabinoid
Phytocannabinoids ("plant cannabinoids") are made by the cannabis plant. THC and CBD are the main examples.
Endocannabinoids ("internal cannabinoids") are made by the body itself. The two best studied are anandamide and 2-arachidonoyl glycerol, usually shortened to 2-AG (Lu & Mackie, 2016).
Pharmaceutical cannabinoids are medicines with a defined content, such as nabilone, dronabinol and nabiximols, which the National Academies of Sciences, Engineering, and Medicine reviewed as treatments (NASEM, 2017).
Synthetic cannabinoid receptor agonists are laboratory chemicals designed to act on cannabinoid receptors. Section 4 describes them as new psychoactive substances.
Canadian law follows the plant. As of 2026, Schedule 1 of the Cannabis Act includes any part of a cannabis plant, including the phytocannabinoids it produces, and any substance identical to a phytocannabinoid however it was obtained (Cannabis Act, 2018). CBD extracted from hemp and CBD made in a laboratory are therefore both regulated as cannabis. Synthetic cannabinoid receptor agonists that are not identical to a phytocannabinoid are listed instead in Schedule II of the Controlled Drugs and Substances Act (CDSA, 1996), a distinction Section 4 returns to.
1.2 The Endocannabinoid System
Cannabis borrows a signalling system that the body already uses. The endocannabinoid system is a widespread system that modulates the activity of nerve cells, with roles in brain development, in the strengthening and weakening of connections between neurons (synaptic plasticity) and in the response to injury (Lu & Mackie, 2016). It has three parts: cannabinoid receptors, the endocannabinoids that activate them, and the enzymes that make and break down those endocannabinoids.
1.2.1 Discovery of the Receptor and Its Natural Ligands
For much of the twentieth century, researchers suspected that cannabinoids acted by dissolving into and disrupting cell membranes in a non-specific way, as some anaesthetics were thought to do. In 1990, Lisa Matsuda and colleagues at the United States National Institute of Mental Health reported the cloning of a receptor that responded more strongly to psychoactive cannabinoids than to non-psychoactive ones and whose messenger RNA was found in brain regions known to bind cannabinoids (Matsuda et al., 1990). This receptor is now called the CB1 receptor. In 1992, a team that included Raphael Mechoulam in Jerusalem isolated arachidonylethanolamide from pig brain, showed that it bound to the cannabinoid receptor and named it anandamide (Devane et al., 1992).
1.2.2 CB1 and CB2 Receptors
Both cannabinoid receptors belong to the family of G protein-coupled receptors that Lesson 3 introduced. When activated, they reduce the activity of the cell, for example by inhibiting certain calcium channels (Lu & Mackie, 2016). The two receptors differ in where they are found:
- CB1 receptors are abundant in the central nervous system, particularly in the cortex, the basal ganglia, the hippocampus and the cerebellum, and most of them sit on axon terminals, the parts of a neuron that release neurotransmitter (Lu & Mackie, 2016).
- CB2 receptors are expressed at much lower levels in the brain, mainly on microglia (the brain's immune cells) and blood vessels, and their expression can increase as much as 100-fold after tissue injury or during inflammation (Lu & Mackie, 2016).
These CB1-rich regions serve memory (the hippocampus), movement and coordination (the basal ganglia and cerebellum) and thinking and perception (the cortex), which is consistent with the effects of acute intoxication that Section 2 describes. 3-D Model 7.1 shows this distribution on a map of CB1 receptors made with positron emission tomography (PET), a scanning method that measures where an injected radioactive tracer binds, and its last step shows the serotonin 5-HT2A receptor on which the classic psychedelics of Section 3 act.
1.2.3 Signals That Travel Backward
Endocannabinoids work differently from the classical neurotransmitters that earlier lessons described. Dopamine, glutamate and GABA are made in advance and stored in small packages (vesicles) in the sending neuron. Endocannabinoids are made on demand from fats in the cell membrane, in one or two rapid enzymatic steps, when a neuron is strongly active (Lu & Mackie, 2016). They are also made in the receiving (postsynaptic) neuron and travel backward across the synapse to CB1 receptors on the sending (presynaptic) terminal. This is called retrograde signalling. When 2-AG activates presynaptic CB1 receptors, the terminal releases less neurotransmitter for a time, so the endocannabinoid system acts as a brake that a busy neuron can apply to its own inputs. Figure 7.1 shows the sequence.
The two endocannabinoids are not equally strong activators. In the vocabulary of Lesson 3, 2-AG is a high-efficacy agonist at both CB1 and CB2 receptors, while anandamide is a low-efficacy agonist at CB1 receptors (Lu & Mackie, 2016). Efficacy describes how fully a bound molecule switches on a receptor, a property separate from how tightly it binds, and it matters again for THC and for the synthetic cannabinoids of Section 4. 3-D Model 7.2 lets you step through retrograde signalling at a synapse that uses glutamate, from the making of 2-AG to its breakdown, and then compare the weak activation of CB1 receptors by THC with the stronger activation by synthetic cannabinoids.
1.3 THC and CBD
THC and CBD come from the same plant, but they have different effects and they are present in very different proportions in different products. Table 7.1 compares them on the features this lesson uses.
Table 7.1. THC and CBD compared on their effects, their receptor actions, recent market trends and their legal status in Canada.
| Feature | THC (delta-9-tetrahydrocannabinol) | CBD (cannabidiol) |
|---|---|---|
| Intoxication | Produces the high and intoxication of cannabis (Health Canada, 2023a). | Does not produce a high or intoxication (Health Canada, 2023a). |
| Action at CB1 | A low-efficacy (partial) agonist, like anandamide (Lu & Mackie, 2016). | Its actions are less well characterized; it does not reproduce THC's intoxication. |
| Relation to harm | Higher doses can increase the risk of harm (Freeman & Lorenzetti, 2020). | May offset some of THC's adverse effects on mental and cognitive outcomes, though not all, and may itself contribute to some driving-related impairment (Fischer et al., 2022). |
| Trend in markets, 1970 to 2017 | Rose in herbal cannabis and resin (Freeman et al., 2021). | No evidence of change (Freeman et al., 2021). |
| Legal status in Canada (2026) | Both are phytocannabinoids and are regulated as cannabis under Schedule 1 of the Cannabis Act (current to 30 September 2026). | |
1.3.1 THC as a Partial Agonist
Lesson 3 explained that a partial agonist activates a receptor less fully than a full agonist does, and Lu and Mackie (2016) describe THC, like anandamide, as a low-efficacy agonist at CB1 receptors. They add that where receptors are few, THC may also partly block the signal of 2-AG, the body's own high-efficacy messenger at the same receptors. Two human observations illustrate what this means. When volunteers who had used cannabis were given rimonabant, a drug that blocks CB1 receptors, even repeated high doses only modestly reduced their reported "high", although rimonabant substantially suppressed the rapid heart rate that cannabis causes. Lu and Mackie take this modest effect as evidence that THC both weakly activates CB1 receptors and partly blocks the body's own signalling there. Rimonabant also did not set off withdrawal in people who had taken moderate doses of THC under supervision, although it did in rodents given high doses of THC for a long time (Lu & Mackie, 2016). These findings contrast with opioids, where Lesson 5 showed that the antagonist naloxone rapidly reverses the effects of morphine. Cannabis intoxication is therefore harder to switch off with a single receptor blocker, and it depends on dose, on how many receptors are available and on how often the person uses cannabis.
1.3.2 Timing of Effects
As Lesson 3 showed for other drugs, route shapes timing. Smoking and vaping cannabis produce similarly rapid onset, while eating or drinking it delays the onset and lengthens the effects, which can lead people to consume more than they intended (Fischer et al., 2022). The 2022 Lower-Risk Cannabis Use Guidelines (Section 2) estimate that inhaled cannabis may impair driving skills for about 6 to 8 hours and edibles for 8 to 12 hours, with wide variation between people and occasions (Fischer et al., 2022).
2.1 Cannabis Products in Canada
The cannabis sold in Canada's legal market is a range of products with very different THC content: Health Canada (2023a) lists dried or fresh plant material, oils, extracts, edibles and topicals. The regulations that took effect on 17 October 2019 set limits on THC per package or per unit for the newer product classes (Health Canada, 2019), and the same limits remained in the regulations as of 30 September 2026 (Cannabis Regulations, 2018):
- a package of edible cannabis may contain at most 10 milligrams (mg) of THC;
- a unit of a cannabis extract, such as a capsule, may contain at most 10 mg of THC, and a package of extract at most 1,000 mg;
- a package of a cannabis topical may contain at most 1,000 mg of THC.
The same backgrounder describes restrictions on ingredients and packaging intended to make edible products less appealing to young people (Health Canada, 2019). The flip cards below describe each form.
Tassone and colleagues (2023) catalogued every product offered by the Ontario Cannabis Store, the public agency that, at the time, ran the province's only online store and supplied all its licensed retail stores, between 19 January and 23 March 2022. They recorded 1,771 products. Table 7.2 summarizes the THC content of the main product types.
Table 7.2. THC content of selected products offered by the Ontario Cannabis Store, January to March 2022 (Tassone et al., 2023).
| Product type | Products listed | Mean THC | Share with 20% THC or more |
|---|---|---|---|
| Dried flower | 508 | 22% by weight | 94% |
| Pre-rolls | 305 | 21% by weight | 93% |
| Thread vape cartridges | 230 | 74% by weight | 96% |
| Hash | 31 | 44% by weight | 100% |
| Resin | 50 | 71% by weight | 100% |
| Shatter and wax | 31 | 75% by weight | 100% |
| Soft chews | 104 | 4 mg per unit | Not applicable |
| Beverages | 106 | 4 mg per unit | Not applicable |
The authors noted that high-THC options dominated inhaled products, that balanced or high-CBD options were less available, and that the average THC of the legal market was higher than the 10 to 13 percent Health Canada reported before legalization (Tassone et al., 2023). Store catalogues such as this one give THC as a percentage of weight. As of 2026, Canadian product labels instead state it as a concentration in milligrams per gram for dried cannabis and for extracts such as vape cartridges and shatter, so that 22 percent appears as 220 mg/g, and as an amount in milligrams per unit or per package for edibles and capsules (Cannabis Regulations, 2018; Health Canada, 2026a). The calculator later in this section helps to bridge these forms.
3.1 Rising THC Potency
In drug science, the potency of a cannabis product usually means its concentration of THC. (Lesson 3 used potency in the stricter pharmacological sense of the dose needed for a given effect; for cannabis products, the everyday term refers to concentration.) This part reviews the evidence that THC concentrations have risen and explains how to put a number on the THC in a product.
A systematic review and meta-analysis by Tom Freeman and colleagues (2021) brought this evidence together; the authors searched for studies reporting mean THC and CBD concentrations over at least three annual time points. They found 12 eligible studies from the United States, the United Kingdom, the Netherlands, France, Denmark, Italy and New Zealand. Across 66,747 samples of herbal cannabis from eight studies between 1970 and 2017, THC concentration increased by 0.29 percentage points per year (95 percent confidence interval 0.11 to 0.47). In cannabis resin, it increased by 0.57 percentage points per year (0.10 to 1.03) across 17,371 samples from 1975 to 2017. There was no evidence that CBD concentrations changed in either product. The authors attributed the rise in herbal cannabis to the growing market share of sinsemilla, the seedless, high-THC flowering tops of unpollinated female plants, relative to traditional lower-THC herbal cannabis (Freeman et al., 2021).
A rise of 0.29 percentage points a year sounds small. If that average rate held over the 47 years from 1970 to 2017, it would add up to an increase of about 13.6 percentage points (0.29 × 47 = 13.6), several times the average of about 3 percent that Health Canada gives for dried cannabis in the 1980s. Health Canada (2023a) gives a similar picture for Canada, stating, on a page last updated in November 2023, that the THC content of dried cannabis rose from an average of 3 percent in the 1980s to around 15 percent.
A long-running single series comes from the University of Mississippi's analysis of cannabis seized by the United States Drug Enforcement Administration. ElSohly and colleagues (2016) analyzed 38,681 samples received between 1995 and 2014. The mean THC concentration of the plant material rose from 3.96 percent in 1995 to 11.84 percent in 2014, while CBD fell from 0.28 percent to 0.15 percent, so the ratio of THC to CBD rose from about 14 to 1 to about 79 to 1. Figure 7.2 plots that series and adds, as a separate marker, the mean THC of legal dried flower in Ontario in early 2022.
These sources have limits that a careful reader should keep in mind. Seized samples reflect what police happened to seize and send for testing, and Freeman and colleagues (2021) rated most of their included studies as using non-random sampling. Catalogue data describe what is offered for sale, which may differ from what people buy. Labels can also be inaccurate. Even so, the direction of change was the same across the countries and decades that Freeman and colleagues (2021) reviewed, so the evidence that THC concentrations have risen is consistent across studies, although the size of the rise varies with the market and the method: in those studies, products contained far more THC, relative to CBD, than a generation earlier.
3.2 Putting a Number on THC: The Standard THC Unit
Lesson 4 showed how the Canadian standard drink puts different beverages on a common scale. Freeman and Lorenzetti (2020) argued that cannabis needs a similar unit, because a gram of one product can contain many times more THC than a gram of another, and they proposed a "standard THC unit" fixed at 5 mg of THC for all cannabis products and methods of administration. The box below sets out the reasoning.
Why count THC in milligrams?
Freeman and Lorenzetti (2020) noted that earlier proposals measured cannabis in joints or grams. Counting joints misses pipes, vaporizers, edibles and concentrates, and counting grams ignores how much THC each gram contains. A unit based on milligrams of THC, like an alcohol unit based on grams of ethanol, works across all products.
Three cautions apply. The unit counts the THC in a product, and Lesson 3 showed that the amount reaching the bloodstream depends on route and bioavailability. The proposal is also a research and public health tool; it is not part of Canada's labelling rules, which, as of 2026, state THC in milligrams per gram, per unit or per package. Finally, the unit measures quantity and sets no safe limit: the 2022 Lower-Risk Cannabis Use Guidelines (Section 2) give no limit in standard units, and their authors noted in 2022 that no empirically defined potency standards or risk thresholds existed yet (Fischer et al., 2022).
The arithmetic needs only one conversion. One percent of one gram is 10 mg, so the THC in a dried or concentrated product equals its mass in grams multiplied by its THC percentage and by 10.
Worked Example 7.1. One gram of dried flower at 22 percent THC, the Ontario mean in Table 7.2, contains 1 × 22 × 10 = 220 mg of THC, or 220 ÷ 5 = 44 standard THC units. One gram of plant material at 3.96 percent, the mean of seized cannabis in the United States in 1995, contains 39.6 mg, or about 7.9 units. The same mass therefore carries more than five times as much THC (220 ÷ 39.6 = 5.6). A package of edible cannabis at the legal maximum of 10 mg contains 2 units. Use Calculator 7.1 to check these figures and to compare other products.
Using the means in Table 7.2 and Calculator 7.1, find the THC content of 0.5 g of dried flower (22 percent) and of 0.5 g of the contents of a thread vape cartridge (74 percent), in milligrams and in standard THC units. Then explain in two sentences why a THC concentration, given as a percentage or in milligrams per gram, can be hard to compare with an edible label that reports milligrams.
Suggested answer: 0.5 g at 22 percent contains 110 mg (22 units), and 0.5 g at 74 percent contains 370 mg (74 units), more than three times as much. A percentage or a value in milligrams per gram describes concentration, while an edible label in milligrams states an amount; converting both to milligrams of THC in the product, or to standard units, puts them on one scale.
This section has described what cannabis contains and how its main constituent acts. THC works on a receptor system that shapes memory, movement and thinking, and products in today's markets carry far more THC than those of earlier decades. Whether these facts translate into harm is an empirical question, and Section 2 examines how strong the evidence is for each of the health effects attributed to cannabis.
Key Takeaways
- 1.1 Cannabis contains more than 100 cannabinoids, including THC, which Health Canada names as the most researched, and CBD. Phytocannabinoids come from the plant, endocannabinoids are made by the body, pharmaceutical cannabinoids are medicines of defined content, and synthetic cannabinoid receptor agonists are laboratory chemicals. As of 2026, Schedule 1 of the Cannabis Act covers the plant, its phytocannabinoids and identical substances however they are obtained, while synthetic agonists that are not identical to a phytocannabinoid fall under Schedule II of the CDSA.
- 1.2 The endocannabinoid system consists of cannabinoid receptors, the endocannabinoids anandamide and 2-AG, and the enzymes that make and break them down. CB1 receptors are abundant on axon terminals in the cortex, basal ganglia, hippocampus and cerebellum, while CB2 receptors sit mainly on immune cells and blood vessels. Endocannabinoids are made on demand by an active receiving neuron and travel backward across the synapse to reduce neurotransmitter release.
- 1.3 THC is a low-efficacy (partial) agonist at CB1 receptors that produces the high of cannabis, while CBD does not produce intoxication and may offset some, but not all, of THC's adverse effects. Because a CB1 blocker only modestly reduces the high, cannabis intoxication is harder to switch off with a single receptor blocker than the effects of opioids. Inhaled cannabis acts quickly, while edibles act later and longer, which can lead people to consume more than they intended.
- 2.1 As of 2026, Canada's legal market sells dried flower and pre-rolls, vape cartridges, extracts, edibles and beverages, and topicals, with limits of 10 mg of THC per package of edible cannabis and per unit of extract. In the Ontario Cannabis Store's catalogue in early 2022, high-THC options dominated inhaled products, balanced or high-CBD options were less available, and dried flower averaged 22 percent THC.
- 3.1 THC concentrations rose by about 0.29 percentage points a year in herbal cannabis and 0.57 in resin from 1970 to 2017, with no evidence of change in CBD, and cannabis seized in the United States rose from 3.96 to 11.84 percent THC between 1995 and 2014. Seized and catalogue samples are not random and labels can be inaccurate, but the direction of change was the same across the countries and decades that Freeman and colleagues reviewed, although the size of the rise varies with the market and the method.
- 3.2 The THC in a dried or concentrated product equals its mass in grams multiplied by its THC percentage and by 10, and dividing by 5 gives standard THC units. One gram of flower at 22 percent contains 220 mg, or 44 units, while an edible package at the 10 mg limit contains 2 units. The unit counts the THC in the product rather than the amount absorbed, and it sets no safe limit.
Reflection
Health Canada (2023a), on a page last updated in 2023, states that the average THC content of dried cannabis rose from about 3 percent in the 1980s to around 15 percent. In the Ontario Cannabis Store's catalogue in early 2022, dried flower averaged 22 percent THC and thread vape cartridges 74 percent, while, as of 2026, Canada's regulations limit a package of edible cannabis to 10 mg of THC. Freeman and Lorenzetti (2020) proposed a standard THC unit of 5 mg. Explain why THC concentration matters for describing how much cannabis a person is exposed to, using at least one calculation, and give one argument for and one argument against requiring every cannabis label in Canada to state its THC content in standard units.
Concentration matters because the same amount of product can carry very different amounts of THC. One gram of dried flower at 22 percent contains 1 × 22 × 10 = 220 mg of THC, or 44 standard units, while one gram at the 3 percent average Health Canada gives for the 1980s contains 30 mg, or 6 units. Someone who smokes the same number of joints today as a person did in the 1980s may therefore be exposed to several times as much THC. A vape cartridge at 74 percent concentrates even more THC into a small volume, and an edible package at the 10 mg limit contains only 2 units.
An argument for standard units on every label is that they put flower, extracts and edibles on one scale, much as standard drinks do for alcohol, which would let people and researchers compare products and would support guidelines based on quantity. An argument against is that units count the THC in the product, and Lesson 3 showed that the amount absorbed depends on route and bioavailability; a smoked unit and an eaten unit do not produce the same exposure, so the label could give a false sense of precision unless it is explained.
Minimum 20 characters required.
Question 1: Where on neurons are most CB1 receptors located, according to Lu and Mackie (2016)?
Question 2: Which observation best fits the description of THC as a low-efficacy (partial) agonist at CB1 receptors?
Question 3: A 1 g sample of dried flower contains 22 percent THC. Using a standard THC unit of 5 mg, how many units does it contain?
Question 4: What did the meta-analysis by Freeman et al. (2021) find about cannabis concentrations between 1970 and 2017?
Health Effects of Cannabis by Strength of Evidence
Introduction
Public claims about cannabis range from a cure for many illnesses to a cause of serious mental illness, and both kinds are often stated with more confidence than the evidence allows. Section 1 showed that THC acts on a receptor system involved in memory, movement and thinking, and that products have become much stronger. Here the question is which health effects are well established, which are uncertain and which have little support; two outcomes, psychosis and cannabis use disorder, receive closer attention, followed by Canada's guidelines for lower-risk use. HSCI 402 (Substance Use, Addiction and Public Health) develops screening, diagnosis and treatment, and Lesson 12 examines what changed after legalization.
Learning Objectives
- 4.1 Explain the weight-of-evidence categories used by the National Academies of Sciences, Engineering, and Medicine (NASEM, 2017) and sort common claims about cannabis into them.
- 4.2 Interpret the evidence linking cannabis use, especially daily use of high-potency products, to psychotic disorders, including what a population attributable fraction does and does not show.
- 4.3 Describe cannabis use disorder and compare published estimates of how common it is among people who use cannabis.
- 5.1 Summarize Canada's Lower-Risk Cannabis Use Guidelines and explain how the 2022 update changed the 2017 recommendations.
4.1 Grading Claims: The NASEM Review
In 2016, the National Academies of Sciences, Engineering, and Medicine in the United States convened a committee of 16 experts to review the evidence on the health effects of cannabis published since a 1999 report. The committee considered more than 10,700 abstracts, gave priority to systematic reviews published since 2011 and to high-quality primary research, and reached nearly 100 research conclusions (NASEM, 2017). To make those conclusions comparable, it placed each one in a standard weight-of-evidence category.
4.1.1 Five Categories of Evidence
Table 7.3 summarizes the categories. Two features of the wording matter. For therapeutic effects, the committee asked whether cannabis or cannabinoids are an effective treatment. For other health effects, it asked whether cannabis use is statistically associated with the outcome, a deliberately cautious phrase, since most of that evidence comes from observational studies of smoked cannabis in which chance, bias and confounding can never be fully excluded (NASEM, 2017).
Table 7.3. The NASEM (2017) weight-of-evidence categories, summarized from the report's Box S-3.
| Category | Strength of support | Typical body of evidence | How firm the conclusion is |
|---|---|---|---|
| Conclusive | Strong evidence from randomized controlled trials | Many supportive findings from good-quality studies, with no credible opposing findings | Firm; chance, bias and confounding can be ruled out with reasonable confidence |
| Substantial | Strong evidence | Several supportive findings from good-quality studies, with very few or no credible opposing findings | Firm, but minor limitations cannot be ruled out |
| Moderate | Some evidence | Several supportive findings from good- to fair-quality studies | A general conclusion, with limitations that cannot be ruled out |
| Limited | Weak evidence | Supportive findings from fair-quality studies, or mixed findings that mostly favour one conclusion | A conclusion with significant uncertainty |
| No or insufficient | No or insufficient evidence | Mixed findings, a single poor study, or an outcome that has not been studied | No conclusion is possible |
A category can also describe evidence against an effect. The committee found moderate evidence of no statistical association between smoking cannabis and lung cancer, for example (NASEM, 2017). In its summary, the report grouped the conclusive and substantial categories together when listing therapeutic effects, and none of its conclusions about harms reached the conclusive level. Interactive 7.1 groups them in the same way. Try to place each claim before reading the summary that follows it.
For each claim, choose the weight-of-evidence category that NASEM (2017) assigned, then select Check my sorting. The feedback gives the committee's conclusion number and wording.
4.1.2 The Evidence by Area
The tabs below summarize selected conclusions from the report's annex, with each conclusion's category in brackets. In the 2017 ratings, the pattern was strong evidence for a few therapeutic uses and a few harms, and limited or insufficient evidence for many popular claims in both directions (NASEM, 2017).
- Cannabis for chronic pain in adults; oral cannabinoids for nausea and vomiting caused by chemotherapy; oral cannabinoids for patient-reported spasticity in multiple sclerosis (conclusive or substantial).
- Cannabinoids, primarily nabiximols, for short-term sleep outcomes in people with sleep disturbance linked to sleep apnea, fibromyalgia, chronic pain or multiple sclerosis (moderate).
- Appetite and weight loss in HIV/AIDS, symptoms of Tourette syndrome, and anxiety during a public speaking test in social anxiety disorder (limited).
- Cancers including glioma, epilepsy, and achieving abstinence from other addictive substances (no or insufficient evidence, as of 2017; for epilepsy, see the 2023 review below).
- Worse respiratory symptoms and more frequent chronic bronchitis with long-term smoking; increased risk of motor vehicle crashes; lower birth weight with maternal cannabis smoking (substantial).
- Overdose injuries, including respiratory distress, among children in United States states where cannabis is legal (moderate).
- No association with lung cancer or with head and neck cancers (moderate evidence of no association).
- Triggering of acute myocardial infarction, and ischemic stroke or subarachnoid hemorrhage (limited).
- All-cause mortality, and death due to cannabis overdose (no or insufficient evidence).
- Development of schizophrenia or other psychoses, with the highest risk among the most frequent users (substantial).
- Impaired learning, memory and attention with acute use; more symptoms of mania in bipolar disorder with regular use; a small increased risk of depressive disorders; more social anxiety disorder with regular use (moderate).
- Impaired academic achievement and education outcomes; development of bipolar disorder among regular or daily users (limited).
- Increasing frequency of use is associated with progression to problem cannabis use, and starting at an earlier age is a risk factor (substantial).
- Major depressive disorder is a risk factor for problem cannabis use (moderate).
- Cannabis use is associated with developing dependence on other substances, including alcohol and tobacco (moderate).
The NASEM report describes evidence published up to 2016, and research in this area is evolving. A 2023 umbrella review of meta-analyses, for example, rated the observational evidence linking cannabis use with psychosis, car crashes and low birth weight as convincing, and found from randomized trials that cannabidiol is effective in people with epilepsy, an outcome the 2017 report had rated as having insufficient evidence (Solmi et al., 2023). The studies below add further detail on psychosis.
4.2 Cannabis and Psychosis
Psychosis is a state in which a person loses contact with shared reality, often through hallucinations or delusions; schizophrenia is the best-known disorder in which psychosis recurs. NASEM (2017) found substantial evidence of an association between cannabis use and the development of schizophrenia or other psychoses, with the highest risk among the most frequent users. Lu and Mackie (2016) note that acute use of cannabis or synthetic cannabinoids can produce psychotic symptoms in susceptible people. A 2022 consensus paper by a task force of the World Federation of Societies of Biological Psychiatry concluded that converging lines of evidence link cannabis to a higher risk of psychosis, and that the risk rises with dose and with earlier exposure, while alternative explanations such as reverse causation and confounding cannot be conclusively excluded (D'Souza et al., 2022). Case 7.1 describes one large study that illustrates the dose-response pattern and asked how much psychosis cannabis might account for.
Marta Di Forti and colleagues (2019) recruited 901 adults aged 18 to 64 who came to psychiatric services with a first episode of psychosis at 11 sites in Europe and Brazil between May 2010 and April 2015, and 1,237 population controls from the same areas. Participants reported how often they used cannabis and which type. The researchers classified each type as high potency (10 percent THC or more) or low potency using national data; in London, skunk-like cannabis averaging 14 percent THC made up 94 percent of the street market, and in the Netherlands THC reached up to 67 percent in Dutch-produced hash (Nederhasj) and 22 percent in Dutch-grown herbal cannabis (Nederwiet).
After adjusting for age, sex, ethnicity, education, employment and use of tobacco and other drugs, daily cannabis use was associated with about 3.2 times the odds of psychotic disorder compared with never use, and daily use of high-potency cannabis with about 4.8 times the odds (Di Forti et al., 2019).
An odds ratio compares the odds of the outcome in people with and without an exposure; a value of 1 means no difference. Figure 7.3 shows the main odds ratios with their 95 percent confidence intervals. The odds rise with frequency of use and are highest for daily use of high-potency products.
4.2.1 How Much Psychosis Might Cannabis Account For?
A population attributable fraction (PAF) estimates the proportion of cases in a population that would not have occurred if an exposure had been absent, assuming that the exposure causes the outcome. It depends on two things: how strongly the exposure is associated with the outcome and how common the exposure is among the cases. For case-control studies, a widely used formula needs only the share of cases who were exposed and the odds ratio.
Worked Example 7.2. In Amsterdam, 69.6 percent of the people with first-episode psychosis had used high-potency cannabis, and the adjusted odds ratio was 3.6. The PAF is 0.696 × (3.6 − 1) ÷ 3.6 = 0.696 × 0.722 = 0.503, or 50.3 percent, the value the authors reported (Di Forti et al., 2019). In London, 51.5 percent of cases had used high-potency cannabis and the odds ratio was 2.4, giving 0.515 × 1.4 ÷ 2.4 = 30.0 percent; the published value of 30.3 percent was computed with statistical software from the adjusted model, so it differs slightly. Table 7.4 lists selected results, and Calculator 7.2 lets you reproduce them.
Table 7.4. Population attributable fractions for psychotic disorder in selected EU-GEI sites, assuming causality (Di Forti et al., 2019, Table 3).
| Site and exposure | Adjusted OR | Controls exposed | Cases exposed | Published PAF |
|---|---|---|---|---|
| All 11 sites, daily use | 3.2 | 6.8% | 29.5% | 20.4% |
| All 11 sites, high-potency use | 1.6 | 19.1% | 35.1% | 12.2% |
| London, high-potency use | 2.4 | 26.0% | 51.5% | 30.3% |
| Amsterdam, high-potency use | 3.6 | 54.0% | 69.6% | 50.3% |
| Amsterdam, daily use | 7.1 | 13.1% | 51.0% | 43.8% |
| Bologna, high-potency use | 1.2 | 8.7% | 11.1% | 1.9% |
Note. Equation 7.2 reproduces the other published values in this table to within rounding. For the all-sites high-potency row it gives 0.351 × (1.6 − 1) ÷ 1.6 = 13.2 percent, compared with the published 12.2 percent, and rounding of the odds ratio does not account for the gap, since any odds ratio that rounds to 1.6 gives about 12.5 to 13.8 percent. As with London, the authors computed each fraction with statistical software from the fully adjusted model, and the paper does not report the details needed to trace the difference (Di Forti et al., 2019).
The study also linked these patterns to differences between places. Across the 11 sites, the incidence of psychotic disorder was correlated with the prevalence of daily use among controls (r = 0.8) and with the prevalence of high-potency use (r = 0.7). Assuming causality, the authors estimated that removing high-potency cannabis would lower the incidence of psychotic disorder in Amsterdam from 37.9 to 18.8 cases per 100,000 person-years, and in London from 45.7 to 31.9 (Di Forti et al., 2019).
4.2.2 Weighing the Evidence
These results are consistent with NASEM's conclusion, and the box below sets out the questions the design leaves open. Each is a reason for caution about the size of the effect, and none on its own explains away a dose-response gradient seen across sites.
Reading an attributable fraction critically
Causality is assumed. A PAF answers "how many cases if the association were causal". If part of the association reflects confounding (a third factor such as genetic risk or adversity that raises both cannabis use and psychosis) or reverse causation (early symptoms leading people to use cannabis), the true fraction is smaller.
Exposure was self-reported. Cannabis use was not confirmed with urine, blood or hair tests, although the authors cite evidence that people report frequency and type of use reliably (Di Forti et al., 2019).
Potency was inferred. THC content was assigned from the type of cannabis and from national market data, so the potency of what each person actually used is unknown.
Canadian researchers have also used Ontario health records to follow how often new diagnoses of schizophrenia are associated with cannabis use disorder before and after legalization (Myran et al., 2025). HSCI 214 (Perspectives on Mental Health and Illness) develops the epidemiology of psychotic disorders, including the evidence on cannabis, and Lesson 12 examines how cannabis legalization has been evaluated.
4.3 Cannabis Use Disorder
A substantial minority of people who use cannabis develop problems with it. Cannabis use disorder is the clinical term for a pattern of use that causes distress or impairment. The 2022 Lower-Risk Cannabis Use Guidelines describe it as characterized by symptoms such as craving, withdrawal, neglect of essential obligations and a limited capacity to control or reduce use (Fischer et al., 2022). HSCI 402 develops the diagnostic criteria of the DSM-5-TR and treatment, and Lesson 8 explains the brain changes behind tolerance and withdrawal.
Published estimates differ, and the differences are instructive. Health Canada's public information page, last updated in October 2025, states that about 1 in 3 people who use cannabis will develop a problem with their use, that about 1 in 11 (9 percent) will develop an addiction, that this rises to about 1 in 6 (17 percent) for people who start as teenagers, and that the risk is 25 to 50 percent among people who use cannabis daily (Health Canada, 2025a). A meta-analysis by Leung and colleagues (2020), which pooled 21 studies identified in searches up to 2020, found that 22 percent of people who used cannabis met criteria for cannabis use disorder, 13 percent for the older diagnosis of cannabis dependence, and that in cohort studies the risk of dependence rose to 33 percent among young people who used weekly or daily.
The figures fit together once definitions are compared: "addiction" and "dependence" describe the more severe end of the spectrum, while cannabis use disorder under current criteria also includes mild cases and is therefore more common. Both sources point to the two risk factors NASEM (2017) rated as substantial, more frequent use and an earlier start, and the 2022 guidelines add high-potency products (Fischer et al., 2022).
5.1 Canada's Lower-Risk Cannabis Use Guidelines
Lesson 4 described Canada's Guidance on Alcohol and Health as an example of risk communication. The Lower-Risk Cannabis Use Guidelines (LRCUG) play a similar role for cannabis. They state that abstaining is the only reliable way to avoid the risks of cannabis, and for people who do use it they identify the behaviours that most change the risk of harm. The authors describe them as a targeted prevention tool that complements regulation, universal prevention and treatment (Fischer et al., 2017, 2022).
The guidelines were first published in 2011, and Benedikt Fischer and colleagues (2017) published a comprehensive update in the American Journal of Public Health, with 10 recommendations (Fischer et al., 2017, 2022). Health Canada presents these on its website, noting the support of the Canadian Research Initiative in Substance Misuse, funded by the Canadian Institutes of Health Research (Health Canada, 2020). An international team updated them in 2022 with twelve recommendation clusters and three general precautions, each with a stated evidence level, funded by Health Canada's Substance Use and Addiction Program and excluding authors with recent financial interests in cannabis companies (Fischer et al., 2022). Table 7.5 summarizes the 2022 version.
Table 7.5. The 2022 Lower-Risk Cannabis Use Guidelines, summarized, with the evidence level the authors assigned (Fischer et al., 2022).
| Number | Recommendation (summarized) | Evidence level |
|---|---|---|
| Precaution A | No level of use is universally safe; abstaining is the only reliable way to avoid all risk. | Conclusive |
| 1 | Delay starting until after late adolescence or the completion of puberty. | Moderate |
| 2 | Choose low-THC products, or products with a high ratio of CBD to THC, with reliable labelling. | Substantial to moderate |
| 3 | Avoid smoking; every route carries some risk, and edibles act later and longer. | Substantial to moderate |
| 4 | If inhaling, avoid deep inhalation and breath-holding. | Limited |
| 5 | Avoid frequent (daily or near-daily) or intensive use. | Substantial |
| 6 | Where possible, use legal, quality-controlled products. | Limited |
| 7 | If cognitive performance is impaired, pause or substantially reduce use. | Limited |
| 8 | Do not drive or operate machinery while impaired; impairment may last about 6 to 8 hours after inhalation and 8 to 12 hours after edibles, and combining with alcohol multiplies it. | Substantial to moderate |
| 9 | Abstain when trying to conceive, when pregnant and when breastfeeding. | Limited |
| 10 | Be cautious about combining cannabis with other substances or medications. | Moderate to limited |
| 11 | People with a personal or family history of psychosis, mood disorders or substance use disorders, people with cardiovascular risks and older adults should avoid or adjust use. | Moderate to limited |
| 12 | Avoid combining risk factors, such as frequent use of high-potency products from a young age. | Substantial to limited |
| Precautions B and C | Cannabis use disorder may require professional care; avoid use that harms others, including through second-hand smoke. | Substantial (B); substantial to limited (C) |
The 2017 version advised against starting "definitively before the age of 16", against synthetic cannabinoids, and against driving for at least 6 hours after use (Fischer et al., 2017; Health Canada, 2020). The 2022 version moved the age advice to the end of late adolescence or puberty, left synthetic cannabinoids out of its scope, replaced the single driving interval with ranges that depend on the route, added recommendations on legal products, cognitive impairment and combining substances, and widened the groups advised to take particular care to include older adults and people with cardiovascular risks (Fischer et al., 2022).
The 2022 authors cite population data from North America showing that sizable groups of people who use cannabis did not follow key recommendations on mode of use, frequency and driving after use. They argue that the guidelines need audience-tailored dissemination, support from regulation such as product labelling, regular updating and evaluation of their effect (Fischer et al., 2022).
Several recommendations map directly onto NASEM conclusions and the EU-GEI findings. Try it 7.2 asks you to make these links yourself.
For each of the 2022 recommendations numbered 2, 3, 5, 8 and 9 in Table 7.5, name one piece of evidence from this section or from Section 1 that supports it, and state the strength of that evidence.
Suggested answer: Recommendation 2 (low potency) links to the EU-GEI finding that daily high-potency use carried the highest odds of psychosis (Di Forti et al., 2019) and to rising THC concentrations (Freeman et al., 2021). Recommendation 3 (avoid smoking) links to NASEM's substantial evidence on respiratory symptoms and bronchitis. Recommendation 5 (frequency) links to substantial evidence that more frequent use is associated with problem use. Recommendation 8 (driving) links to substantial evidence of an increased risk of motor vehicle crashes. Recommendation 9 (pregnancy) links to substantial evidence of lower birth weight with maternal smoking, and the guideline authors rated the evidence for their recommendation as limited, noting that the effects on offspring are likely small.
Taken together, the reviews in this section support a small number of therapeutic uses of cannabis, and NASEM (2017) found too little evidence to judge any link with death from overdose. In the 2022 guidelines and in a 2022 systematic review of potency, the risks of the most important harms, such as psychosis and use disorder, are highest among people who use frequently, start young or use high-potency products (Fischer et al., 2022; Petrilli et al., 2022). Section 3 turns to drugs that are pharmacologically very different and that are also the subject of renewed medical interest: the classic psychedelics, MDMA and the dissociatives.
Key Takeaways
- 4.1 NASEM (2017) placed nearly 100 conclusions in five weight-of-evidence categories, from conclusive to no or insufficient evidence, and a category can describe evidence against an effect, as with moderate evidence of no association with lung cancer. It found conclusive or substantial evidence for chronic pain, chemotherapy-induced nausea and spasticity in multiple sclerosis, substantial evidence linking cannabis to respiratory symptoms, crashes, lower birth weight, psychosis and problem use, and limited or insufficient evidence for many other claims.
- 4.2 In the EU-GEI study, daily use carried about 3.2 times and daily use of high-potency cannabis about 4.8 times the odds of psychotic disorder compared with never use. Attributable fractions of up to about 50 percent, as in Amsterdam, assume that the association is causal; confounding, reverse causation, self-reported use and inferred potency are reasons for caution about their size.
- 4.3 Cannabis use disorder is marked by craving, withdrawal, neglect of obligations and a limited capacity to control use. Estimates range from about 1 in 11 people who use cannabis developing an addiction to 22 percent meeting criteria for the disorder, because addiction and dependence describe the severe end of a spectrum that also includes mild cases. More frequent use and an earlier start are the risk factors NASEM rated as substantial.
- 5.1 The Lower-Risk Cannabis Use Guidelines state that abstaining is the only reliable way to avoid risk and identify the behaviours that most change risk for people who do use cannabis. The 2022 update gave twelve recommendation clusters with evidence levels, moved the age advice to the end of late adolescence or puberty, replaced the single 6-hour driving interval with ranges by route, left synthetic cannabinoids out of scope, and added advice on legal products, cognitive impairment, combining substances and high-potency products.
Reflection
Suppose a news headline reads: "Strong cannabis causes half of all psychosis in Amsterdam." In the EU-GEI study (Di Forti et al., 2019), 69.6 percent of people with first-episode psychosis in Amsterdam and 54.0 percent of population controls had used high-potency cannabis (10 percent THC or more), the adjusted odds ratio for high-potency use was 3.6, and the authors estimated, assuming causality, a population attributable fraction of 50.3 percent. Self-reported cannabis use was not checked with biological tests, and potency was assigned from the type of cannabis and national market data. Evaluate the headline: state what it gets right, what assumption it hides, and two alternative explanations or limitations that a careful reader should consider.
The headline reports the study's central estimate correctly: using the formula PAF = 0.696 × (3.6 − 1) ÷ 3.6, about 50 percent of new cases of psychosis in Amsterdam would not have occurred if high-potency cannabis had been absent. It is also consistent with NASEM's substantial evidence of an association between cannabis and psychosis, and with a dose-response gradient in which daily use of high-potency cannabis carried the highest odds.
The word "causes" hides the key assumption. A population attributable fraction answers the question "how many cases if the association were causal"; the authors stated that they assumed causality. If part of the association reflects confounding, for example by genetic vulnerability or childhood adversity that raises both cannabis use and psychosis risk, or reverse causation, in which early symptoms lead people to use cannabis, the true fraction would be smaller. Measurement is a second limitation: use was self-reported and potency was inferred from the type of cannabis, so some people may be misclassified. The estimate also applies to Amsterdam's market at the time; a site such as Bologna, where high-potency use was rarer, had a fraction of about 2 percent. A more accurate headline would say "up to half" and "if cannabis is a cause".
Minimum 20 characters required.
Question 1: What did the NASEM (2017) committee conclude about cannabis smoking and lung cancer?
Question 2: In Amsterdam, 69.6 percent of people with first-episode psychosis had used high-potency cannabis, and the adjusted odds ratio was 3.6. Assuming causality, about what share of cases does the formula PAF = pc × (OR − 1) ÷ OR attribute to high-potency cannabis?
Question 3: Health Canada's information page says about 1 in 11 people who use cannabis develop an addiction, while Leung et al. (2020) estimated that 22 percent meet criteria for cannabis use disorder. What best explains the difference?
Question 4: Which change between the 2017 and 2022 Lower-Risk Cannabis Use Guidelines is described correctly?
Classic Psychedelics, MDMA, Ketamine and Other Dissociatives
Introduction
The drugs in this section change perception, emotion and the sense of self more dramatically than any class covered so far, yet several of them rank low on measures of dependence and physical toxicity. That combination has made them the subject of both moral panic, which Lesson 2 traced in the 1960s, and renewed clinical research. Lesson 3 introduced the idea of drug, set and setting, the principle that a drug experience depends on the person's expectations and the context of use as well as on the pharmacology, and that principle is especially visible here. Grouping the drugs into three families by their main receptor target makes their effects, risks and dependence potential easier to compare, both with one another and with the drugs of earlier lessons. HSCI 485, the psychedelic medicine seminar, develops therapeutic use, clinical trials and regulation.
Learning Objectives
- 6.1 Name the three families of drugs once grouped as hallucinogens, their main receptor targets and their status under Canadian law as of 2026.
- 6.2 Name the main classic psychedelics, explain their shared action at the serotonin 5-HT2A receptor, and describe their physiological and psychological risks.
- 6.3 Describe how MDMA acts and its main acute risks, including hyperthermia and hyponatremia.
- 6.4 Describe how ketamine acts and the risks of ketamine and other dissociatives, including ketamine-associated bladder damage.
- 7.1 Compare the dependence potential of these drugs with other drug classes using expert harm ratings.
- 7.2 State the limits of expert harm ratings as a basis for comparing and regulating drugs.
6.1 Three Families of Drugs
The older label "hallucinogens" groups drugs that act in quite different ways. Pharmacology separates them into three families, which Table 7.6 summarizes with their status under Canadian law. Figure 7.4 then shows the main mechanism of each family.
Table 7.6. Classic psychedelics, MDMA and dissociatives: examples, main targets and status under the Controlled Drugs and Substances Act (CDSA) as of 30 September 2026.
| Family | Examples | Main target | Status in Canada (CDSA) |
|---|---|---|---|
| Classic psychedelics (serotonergic hallucinogens) | LSD, psilocybin and psilocin, DMT, mescaline | Agonists or partial agonists at serotonin 5-HT2A receptors (Nichols, 2016) | Schedule III; mescaline is listed "but not peyote" |
| Entactogens | MDMA (ecstasy) | Predominantly enhances the release of serotonin (Liechti, 2015) | Schedule I, listed among the amphetamines |
| Dissociatives | Ketamine, phencyclidine (PCP) | Inhibition of the NMDA type of glutamate receptor (Zanos et al., 2018) | Schedule I, where ketamine is listed under phencyclidine and its derivatives |
Canadian law treats these families differently from one another and from cannabis. Schedule placement affects penalties: as of 2026, under section 4 of the CDSA, the maximum sentence for possession prosecuted by indictment is seven years for a Schedule I substance and three years for a Schedule III substance (CDSA, 1996). Lesson 11 examines how such schedules were built and how they relate to evidence of harm, a question this section's final part takes up.
6.2 Classic Psychedelics
David Nichols (2016) describes psychedelics as powerful psychoactive substances that alter perception and mood and affect many cognitive processes. Their use predates written history; early cultures used them in many social and ritual contexts. Modern scientific interest began with the discovery of LSD at about the same time as serotonin was identified in the brain, which led researchers to ask whether LSD worked through serotonin. By 2016, Nichols could describe a consensus that psychedelics are agonists or partial agonists at serotonin 5-HT2A receptors, with the receptors that matter most sitting on the apical dendrites (the long upward branches) of pyramidal neurons in the cerebral cortex (Nichols, 2016). Brain imaging after intravenous psilocybin or LSD shows decreases in oscillatory power, a measure of synchronized activity, in areas of the default mode network, a set of regions active during self-focused thought (Nichols, 2016).
6.2.1 Physiological Effects and Toxicity
Matthew Johnson and colleagues (2008) reviewed the risks of giving classic psychedelics to research volunteers and concluded that the main safety concerns are psychological. Physical effects during drug action can include dizziness, weakness, tremor, nausea, drowsiness, tingling, blurred vision and dilated pupils, together with moderate increases in pulse and blood pressure, but these are modest even at doses that produce powerful psychological effects. In their review, classic psychedelics generally had relatively low physiological toxicity and had not been shown to cause organ damage or lasting neuropsychological deficits (Johnson et al., 2008). Nichols (2016) likewise describes them as generally considered physiologically safe, and Johnson and colleagues cite Gable's comparisons of acute toxicity, which Lesson 3 introduced, among the evidence for this conclusion.
6.2.2 Psychological Risks
The accordion below summarizes the psychological risks that Johnson and colleagues (2008) identified, starting with the most likely.
The most likely risk is a period of overwhelming anxiety, fear or panic, low mood or paranoia during drug action. Because emotions are intensified, distress in an unprepared person or an uncontrolled setting can lead to erratic and dangerous behaviour, such as leaving a safe place. In the authors' own studies, in which 54 volunteers received high doses of psilocybin, reassurance was sufficient to manage all episodes of acute distress (Johnson et al., 2008).
Psychotic reactions lasting days or months are rare and appear to occur mainly in people who already had a mental illness or a strong vulnerability to one. In a survey of researchers published in 1960, one psychotic reaction lasting more than 48 hours occurred among 1,200 research volunteers given LSD or mescaline, a rate of about 0.8 per 1,000 (Johnson et al., 2008). Research protocols therefore exclude people with a personal or family history of psychotic disorders.
Some people re-experience perceptual changes, such as visual disturbances, after the drug has worn off. When these are distressing or impairing, they may meet criteria for hallucinogen persisting perception disorder (HPPD). Johnson and colleagues described its incidence as unknown but thought to be very low, and as apparently lower in screened research settings than in unsupervised use mixed with other drugs (Johnson et al., 2008).
The research safeguards that Johnson and colleagues (2008) recommend (screening out people with psychotic disorders in themselves or their families, building trust before a session, preparing the person for the range of possible experiences, providing a safe physical setting and having two trained monitors present) are the principles of set and setting from Lesson 3 put into practice. They also explain why findings from controlled research cannot be transferred directly to unsupervised use, where none of these conditions can be assumed.
6.2.3 Dependence
Johnson and colleagues (2008) noted that classic psychedelics are not typically considered drugs of dependence, because they do not produce compulsive drug seeking and animals do not reliably self-administer them, and that they were not associated with a known withdrawal syndrome. A 2024 meta-analysis of adverse events in studies of classic psychedelics found them generally well tolerated in clinical or research settings, although serious adverse events did occur among participants with existing neuropsychiatric disorders, and the quality of monitoring and reporting varied (Hinkle et al., 2024). Most of this safety evidence therefore comes from screened volunteers in research settings, and research on these risks is evolving. Lesson 8 explains how drugs that do produce dependence act on the brain's reward system.
Where medical research fits
Nichols (2016) reports that several double-blind, placebo-controlled phase 2 trials of psilocybin-assisted psychotherapy relieved anxiety and depression in people with cancer-related distress, and that small pilot studies suggested benefit in alcohol and nicotine addiction. Ketamine, in clinical use as an anaesthetic since 1970, also has antidepressant actions (Zanos et al., 2018). On 9 August 2024, the United States Food and Drug Administration declined to approve MDMA-assisted therapy.
These developments belong to HSCI 485, which examines the trials, their limitations and the regulation of psychedelic medicines. A drug's effects and risks in a screened clinical setting differ from those of the same drug bought in an unregulated market.
6.3 MDMA
MDMA (3,4-methylenedioxymethamphetamine) is chemically related both to amphetamine, which Lesson 6 described, and to some classic psychedelics, but it acts through a different mechanism (Johnson et al., 2008). Matthias Liechti (2015) classifies it as an entactogen, a drug that predominantly enhances the release of serotonin. This mechanism also accounts for its main acute dangers. Liechti (2015) lists acute serotonin syndrome (a dangerous excess of serotonin activity), hyperthermia (overheating), seizures and hyponatremia (a dangerously low level of sodium in the blood) as complications of serotonergic drugs. Johnson and colleagues (2008) add that animal studies have shown neurotoxic effects of MDMA at high doses, although it has been judged safe for administration in several controlled human research settings.
Drugs bought in unregulated markets may contain substances other than the one expected, which adds risk; Section 4 describes synthetic cathinones mis-sold as other drugs (EUDA, 2026). Legal status also differs from expert judgements of harm. As of 2026, MDMA is listed in Schedule I, so the maximum penalty for possessing it is the same as for heroin or cocaine, while the expert comparison later in this section rated it lower than both for harm and dependence.
6.4 Ketamine and Other Dissociatives
Ketamine is a mixture of two mirror-image forms of the same molecule and has been in clinical use as an anaesthetic since 1970 (Zanos et al., 2018). Its anaesthetic and pain-relieving effects are attributed mainly to inhibition of NMDA receptors, a type of receptor for glutamate, the brain's main excitatory neurotransmitter, which Lesson 4 introduced in connection with alcohol. Its side effects include dissociative effects (feelings of detachment from the body and surroundings), psychotomimetic effects (symptoms resembling psychosis) and cognitive impairment (Zanos et al., 2018). The body rapidly converts ketamine into several metabolites, some of which are themselves active.
6.4.1 Ketamine and the Bladder
Long-term non-medical use of ketamine is associated with damage to the urinary tract, called ketamine-associated uropathy. Andrade (2025) reviewed this evidence. A meta-analysis of 37 studies of people using ketamine recreationally reported lower urinary tract symptoms in 44 to 77 percent and upper urinary tract disease in 8 to 30 percent, but Andrade argues that these figures are potentially misleading, and more recent studies using different methods of finding cases report lower risks of 2 to 27 percent. The damage appears to arise from contact between the bladder and ketamine and its metabolites in urine, and for people who need long-term ketamine treatment Andrade concludes that higher doses, more frequent dosing, longer treatment and oral use are all potential risk factors (Andrade, 2025). In psychiatric treatment, a systematic review of 27 studies found urological symptoms in 0 to 24.5 percent of patients, mostly mild or moderate, although few studies followed patients long term (Kerr-Gaffney et al., 2025), and Andrade (2025) notes that in 14 randomized trials the prevalence of symptoms differed little between ketamine and comparison groups. The contrast between the two settings mirrors the point made about medical research under 6.2.3: dose, frequency and context shape risk.
6.4.2 New Dissociatives
Ketamine is listed in Canadian law as a derivative of phencyclidine (PCP), and the CDSA schedules PCP together with "its salts, derivatives and analogues" (CDSA, 1996). Chemists have nevertheless produced new dissociatives designed to resemble ketamine. In 2022, Health Canada's Drug Analysis Service identified two of them, fluorodeschloroketamine and fluorexetamine, in Canada for the first time (Health Canada, 2023b). Section 4 explains how such new psychoactive substances are detected.
7.1 Comparing Dependence Potential
Comparing these drugs with alcohol, tobacco, cannabis and opioids requires common criteria. One influential attempt to make that comparison systematically was made by David Nutt and colleagues (2007), who asked experts in the United Kingdom to rate 20 substances on nine parameters of harm grouped into three categories.
- Physical harm: acute harm, chronic harm, and harm from intravenous use.
- Dependence: intensity of pleasure, psychological dependence and physical dependence.
- Social harm: intoxication, other social harms and health-care costs.
Each parameter was scored from 0 (no risk) to 3 (extreme risk). One group of 29 addiction psychiatrists rated 14 substances, and a second, independent group with wider expertise in fields such as pharmacology, epidemiology, law and policing rated all 20 in a series of structured meetings. The two groups produced broadly similar rankings (Nutt et al., 2007). Interactive 7.2 lets you explore the independent group's scores.
Choose a measure. Bars are sorted from highest to lowest; the drugs covered in this lesson are shown in teal and the others in grey. Scores run from 0 (no risk) to 3 (extreme risk). Data: Nutt et al. (2007), Table 3, independent expert group.
The ratings reflect expert judgement in the United Kingdom in 2007 about drugs in the forms usually used there; they are not measurements.
Several patterns stand out. Heroin received the highest dependence score (3.00), followed by cocaine (2.39), tobacco (2.21), street methadone (2.08), barbiturates (2.01) and alcohol (1.93). The drugs of this lesson sit lower: ketamine 1.54, cannabis 1.51, LSD 1.23 and ecstasy 1.13. Their profiles differ in instructive ways. LSD was rated high for pleasure (2.2) but very low for physical dependence (0.3), which fits Johnson and colleagues' (2008) description of a drug that people value but do not use compulsively. Ketamine's mean physical harm score (2.00) was higher than those of alcohol (1.40) and amphetamine (1.81). Khat, alkyl nitrites and ecstasy were in the bottom five places for all three categories of harm (Nutt et al., 2007).
Use Interactive 7.2 to answer three questions. (a) Which drug covered in this lesson has the highest mean physical harm score, and what might explain it? (b) Compare LSD's scores for intensity of pleasure and for physical dependence. What does the gap tell you? (c) Where does cannabis fall on dependence relative to alcohol and tobacco, and how does that compare with the evidence on cannabis use disorder in Section 2?
Suggested answer: (a) Ketamine, with a mean physical harm score of 2.00, driven by acute harm (2.1) and harm from intravenous use (2.1); the evidence on bladder damage with long-term use adds a chronic harm. (b) LSD scored 2.2 for pleasure and 0.3 for physical dependence: people value the experience, but it produces little compulsive use or withdrawal. (c) Cannabis (1.51) is lower than alcohol (1.93) and tobacco (2.21), yet Section 2 showed that about 22 percent of people who use cannabis meet criteria for cannabis use disorder (Leung et al., 2020), so a moderate score still describes a meaningful risk for frequent users.
7.2 The Limits of Expert Ratings
The authors themselves noted several limits. The ratings came from a small number of independent scorers. Harms often cannot be separated from how a drug is used: cannabis is commonly smoked with tobacco, which may have raised its scores, and the harms of GHB, ketamine, ecstasy and alcohol are entangled with the use of other drugs. Every parameter carried equal weight and the scores were simply averaged, which, the authors acknowledged, cannot represent a drug that is extremely harmful in one respect, such as tobacco with its long-term toxicity. Legal status itself changes harm, since the easy availability of alcohol and tobacco affects their scores (Nutt et al., 2007). Averaging also shapes comparisons within the table: alcohol's mean physical harm score (1.40) is lower than both its acute (1.9) and chronic (2.4) scores because the mean also includes the parameter for harm from intravenous use, which does not apply to alcohol. The ratings also describe one country at one time.
Other researchers have criticized the approach more fundamentally. The Toronto pharmacologist Harold Kalant (2010) argued that ranking drugs by harm inevitably requires value judgements about how much each kind of harm matters, so an expert ranking cannot be a purely scientific basis for classifying drugs. Caulkins, Reuter and Coulson (2011) argued that reducing harm to a single measure hides dimensions that cannot be compared, ignores the benefits of some drugs, and treats a drug's harms as fixed when they change with the policy in force. Expert ratings are therefore a useful way to compare drugs on common criteria, and a poor substitute for direct evidence such as the NASEM review or the EU-GEI study in Section 2. The authors drew one lesson for drug policy themselves: the ranking of drugs by harm did not match their legal classification in the United Kingdom at the time (Nutt et al., 2007), a theme Lesson 11 develops.
The classic psychedelics, MDMA and ketamine are established drugs with decades of research behind them. Section 4 turns to new psychoactive substances designed to mimic cannabis, stimulants, MDMA, ketamine and benzodiazepines, about which a 2020 review noted that much is still unknown, especially for the newest (Shafi et al., 2020), and to the early warning systems that try to detect them.
Key Takeaways
- 6.1 Pharmacology separates the older category of hallucinogens into classic psychedelics (LSD, psilocybin, DMT and mescaline), which act at serotonin 5-HT2A receptors; the entactogen MDMA, which enhances serotonin release; and dissociatives such as ketamine and PCP, which inhibit NMDA receptors. Under the CDSA as of 2026, the classic psychedelics are in Schedule III and MDMA and ketamine in Schedule I, which carries a higher maximum penalty for possession.
- 6.2 Classic psychedelics are agonists or partial agonists at 5-HT2A receptors on cortical pyramidal neurons. Safety reviews describe low physiological toxicity and note that they are not typically considered drugs of dependence, but they carry psychological risks (acute distress, rare prolonged psychosis in vulnerable people and hallucinogen persisting perception disorder) that depend strongly on set and setting.
- 6.3 MDMA is an entactogen whose serotonin release explains its acute risks of serotonin syndrome, hyperthermia, seizures and hyponatremia, and drugs bought in unregulated markets may contain other substances. Its Schedule I listing, as of 2026, does not match the 2007 expert ratings that placed it below heroin and cocaine for harm and dependence.
- 6.4 Ketamine, an anaesthetic since 1970, acts mainly by inhibiting NMDA glutamate receptors and has dissociative, psychotomimetic and cognitive side effects. Long-term non-medical use is associated with bladder damage, with dose, frequency and duration as potential risk factors, and new dissociatives such as fluorodeschloroketamine have been designed to resemble it.
- 7.1 In the expert ratings of Nutt et al. (2007), heroin scored highest for dependence (3.00), followed by cocaine, tobacco and alcohol, while ketamine (1.54), cannabis (1.51), LSD (1.23) and ecstasy (1.13) scored lower. LSD was rated high for pleasure but very low for physical dependence, and ketamine's mean physical harm score was higher than alcohol's.
- 7.2 The ratings came from a small group of experts in one country at one time, averaged equally weighted parameters, and could not separate a drug's harms from how it is used or from its legal availability. Critics argue that ranking harms requires value judgements, so expert ratings are useful for comparing drugs on common criteria but are no substitute for direct evidence, and their mismatch with the United Kingdom's legal classification at the time raises questions for drug policy.
Reflection
In the expert ratings of Nutt et al. (2007), on a scale from 0 to 3, the mean dependence scores were 3.00 for heroin, 2.21 for tobacco, 1.93 for alcohol, 1.54 for ketamine, 1.51 for cannabis, 1.23 for LSD and 1.13 for ecstasy. Under Canada's Controlled Drugs and Substances Act as of 2026, LSD and psilocybin are in Schedule III, MDMA and ketamine are in Schedule I (with heroin and cocaine), while alcohol and tobacco are legal, regulated products. Johnson et al. (2008) report that classic psychedelics are not considered drugs of dependence but carry psychological risks, and long-term non-medical ketamine use is associated with bladder damage. Explain two reasons why a drug's legal status may differ from its dependence rating, and one reason why expert ratings alone should not decide how a drug is regulated.
The first reason is that legal status reflects history as well as evidence. Alcohol and tobacco were embedded in Canadian life long before drug laws were written, while possession of LSD became an offence in August 1969, amid the official concern about young people's drug use that Lesson 2 described. The scores show that the legal drugs rate higher on dependence than LSD or ecstasy, so the schedules cannot be read as a ranking of dependence.
The second reason is that dependence is only one kind of harm. Ketamine has a moderate dependence score but a mean physical harm score of 2.00, and long-term use can damage the bladder; LSD rates low on dependence but carries risks of acute distress and, rarely, prolonged psychosis in vulnerable people. A schedule may weigh acute danger, social harm or potential for misuse alongside dependence.
Expert ratings should not decide regulation on their own because they are judgements by a small group in one country and year, they blend the drug with how it is used (cannabis with tobacco, ecstasy with other drugs), and they are shaped by legal status itself, since availability changes harm. Direct evidence such as cohort and case-control studies, and the effects a policy would have on supply and on people who use drugs, also need to be weighed.
Minimum 20 characters required.
Question 1: What receptor action do the classic psychedelics (LSD, psilocybin, DMT and mescaline) share, according to Nichols (2016)?
Question 2: Which statement about ketamine-associated uropathy is best supported by Andrade (2025) and Kerr-Gaffney et al. (2025)?
Question 3: According to Johnson et al. (2008), what is the most likely risk when classic psychedelics are given in research settings?
Question 4: Which statement about the expert ratings of Nutt et al. (2007) is accurate?
New Psychoactive Substances and Drug Early Warning Systems
Content note
This section refers to poisonings and deaths associated with new psychoactive substances in unregulated drug markets. Support is available at any time: the 9-8-8 Suicide Crisis Helpline (call or text 9-8-8), the National Overdose Response Service for people who use drugs alone (1-888-688-6677), and the Hope for Wellness Help Line for Indigenous people across Canada (1-855-242-3310).
Introduction
Sections 1 to 3 described drugs with long histories of use and research. This section turns to substances that appear in drug markets with little history of use or research. Some are designed in laboratories to imitate the effects of cannabis, stimulants, MDMA, ketamine or benzodiazepines while differing enough in chemical structure to fall outside existing drug laws. Others are medicines approved in some countries and sold elsewhere without approval, such as the benzodiazepines phenazepam and etizolam (Brunetti et al., 2021). Lesson 5 showed how nitazene opioids and sedatives such as bromazolam and medetomidine entered Canada's unregulated opioid supply; those cases belong to the wider category of new psychoactive substances, which early warning systems are built to detect. HSCI 402 develops drug checking as a harm reduction service.
Learning Objectives
- 8.1 Define new psychoactive substances, explain why they keep appearing and describe how they are classified.
- 8.2 Describe synthetic cannabinoid receptor agonists, synthetic cathinones and novel benzodiazepines, including how they act and why they cause harm.
- 9.1 Explain how the EU Early Warning System and the UNODC Early Warning Advisory detect, assess and report new substances, and how Canada monitors them.
- 9.2 Interpret Health Canada's data on new psychoactive substances identified for the first time, and identify the limits of those data.
8.1 What Makes a Substance "New"?
The United Nations Office on Drugs and Crime (UNODC) provides the definition used internationally, reproduced in the box below. The UN's text uses the older word "abuse"; this course uses "use" except when quoting.
The UNODC definition of a new psychoactive substance
"A New Psychoactive Substance (NPS) is a substance of abuse, either in a pure form or a preparation, that is not controlled by the 1961 Single Convention on Narcotic Drugs or the 1971 Convention on Psychotropic Substances, but which may pose a 'public health threat'." The UNODC adds that "new" refers to substances that have recently become available, which need not be new inventions, and that substances placed under international control since 2015 continue to be counted as NPS so that trends can be followed over time (UNODC, n.d.).
In its report for 2024, Health Canada used a working definition tied to its laboratory: a substance with potential psychoactive effects that its Drug Analysis Service has identified in Canada for the first time (Health Canada, 2025b). Its report for 2025 extended the definition to substances first identified among seizures by the Canada Border Services Agency (Health Canada, 2026b).
8.1.1 Why New Substances Keep Appearing
New psychoactive substances have been sold under names such as "designer drugs", "legal highs", "herbal highs" and "bath salts". Many are produced by making slight changes to the chemical structure of controlled substances in order to get around drug controls (UNODC, n.d.). When a substance is controlled, a modified version can appear, a pattern that Brunetti and colleagues (2021) describe for designer benzodiazepines as a "cat-and-mouse game" between organized crime and law enforcement. Lesson 11 examines the wider economic forces behind this pattern, including the iron law of prohibition.
The scale is large. The UNODC reports, as of October 2026, that 153 countries and territories from all regions of the world have reported at least one new psychoactive substance to its Early Warning Advisory (UNODC, n.d.). At the end of 2025, the European Union Drugs Agency (EUDA) was monitoring 1,050 new psychoactive substances, 50 of which had been notified for the first time in 2025 (EUDA, 2026).
8.1.2 Why They Are Dangerous
The UNODC (n.d.) summarizes the health problems linked to new psychoactive substances as ranging from seizures to agitation, aggression, acute psychosis and the potential development of dependence, with frequent hospital admissions for severe intoxication. Three features make them especially hard to use safely. Safety data on toxicity are unavailable or very limited for many of them, the purity and composition of products are often unknown, and they are frequently consumed with other drugs; emergency admissions and deaths often involve several substances (UNODC, n.d.). A person may also take a new substance without knowing it, because it was sold as something else.
8.1.3 How New Substances Are Classified
The UNODC classifies new psychoactive substances in two ways (UNODC, n.d.). Structural groups describe chemistry, such as synthetic cannabinoids, synthetic cathinones, benzodiazepines, phenethylamines, tryptamines, fentanyl analogues and nitazenes. Effect groups describe what the substances do: synthetic cannabinoid receptor agonists, classic hallucinogens, stimulants, opioid receptor agonists, sedatives and hypnotics, and dissociatives. The effect groups map closely onto the drug classes of Lessons 4 to 7, which is why the pharmacology of those lessons is the first tool for understanding a substance that has been little studied.
8.2 Three Groups of New Psychoactive Substances
Three structural groups illustrate the range of new psychoactive substances and connect them to earlier sections and lessons. Evidence on all three is still emerging (Shafi et al., 2020): for designer benzodiazepines, for example, Brunetti and colleagues (2021) found that knowledge of their pharmacokinetics and toxicity rested largely on reports of emergency department admissions, impaired driving cases and deaths, and on users' own reports. The tabs describe each group, and Table 7.7 then compares them.
What they are. Synthetic cannabinoid receptor agonists are laboratory chemicals that act on CB1 receptors like THC but are chemically diverse (Liechti, 2015). Cohen and Weinstein (2018) reported that at least 200 had been identified since 2008 in herbal smoking mixtures, sold under names such as "Spice" and "K2" (Health Canada, 2020).
How they act. Section 1 described THC as a low-efficacy agonist. Cohen and Weinstein describe synthetic cannabinoids as, in general, more potent and more efficacious CB1 agonists than THC, many of them full agonists. One early example, JWH-018, binds CB1 receptors with four times the affinity of THC (Cohen & Weinstein, 2018). A product can therefore produce effects of a different kind from cannabis, as well as stronger ones.
Harms. Reported effects include agitation, anxiety, paranoia, high blood pressure, seizures and, rarely, heart attack and kidney failure (Liechti, 2015; Cohen & Weinstein, 2018). The EUDA's European Drug Report 2026 states that in Europe they particularly affect people in prison and people experiencing homelessness, may be mis-sold as or mixed into cannabis products, and increasingly appear in vaping liquids, and that in September 2025 Czechia reported an outbreak of poisonings linked to MDMB-PINACA (EUDA, 2026).
What they are. Synthetic cathinones are stimulants chemically related to amphetamine and to cathinone, which Canadian law lists in Schedule III as of 2026 (CDSA, 1996). Reported seizures and imports of synthetic cathinones in the European Union rose by 11.5 tonnes to 48.5 tonnes in 2024, and four new cathinones were notified in 2025 (EUDA, 2026).
How they act. Lesson 6 described two stimulant mechanisms: blocking the transporters that clear dopamine and noradrenaline, and causing their release. Cathinones show both. Pyrovalerone cathinones such as MDPV block transport, methamphetamine-like cathinones cause release, and MDMA-like cathinones predominantly release serotonin (Liechti, 2015).
Harms. Drugs with strong dopamine effects are highly addictive and can cause prolonged stimulation, insomnia, agitation and psychosis, while serotonergic ones carry the risks of serotonin syndrome, hyperthermia, seizures and hyponatremia described for MDMA in Section 3 (Liechti, 2015). In 2025, N-ethylnorpentedrone (NEP) was increasingly mis-sold in Europe as another cathinone, 3-MMC, leading to unintended use and more poisonings (EUDA, 2026).
What they are. Novel or "designer" benzodiazepines are benzodiazepines that are not approved as medicines where they are sold. Some were developed as medicines elsewhere: phenazepam was developed in the 1970s and is used in countries of the former Soviet Union, and etizolam was introduced in 1983 and is prescribed in a few countries. Others were first identified in seized products, such as flubromazolam, found in Sweden in 2014 in tablets labelled "XANAX", and flualprazolam, reported by Swedish police in 2018 (Brunetti et al., 2021).
How they act. Lesson 4 explained how benzodiazepines enhance the effect of GABA. Novel benzodiazepines act in the same way, but Brunetti and colleagues (2021) note that most have not undergone clinical trials, so their potency, duration and toxicity are poorly known.
Harms. They produce strong sedation and amnesia and raise the risk of breathing suppression and death when combined with other depressants (Brunetti et al., 2021), the interaction Lesson 4 described. Lesson 5 showed that bromazolam and desalkylgidazepam became common in Canada's unregulated opioid supply (CCSA, 2025). Between 2020 and 2021, clonazolam, diclazepam, etizolam, flualprazolam and flubromazolam were added to Schedule IV of the 1971 Convention (Brunetti et al., 2021).
3-D Model 7.3 shows one synthetic cannabinoid, MDMB-FUBINACA, in the measured atomic structure of the CB1 receptor with its G protein attached, and the change in the receptor's shape when an agonist switches it on.
Table 7.7. Three groups of new psychoactive substances compared, with examples first identified in Canada by Health Canada's Drug Analysis Service (Health Canada, 2023b, 2024, 2025b) and their treatment under the CDSA as of 30 September 2026.
| Group | Related established drug | Main action | Examples first identified in Canada | How the CDSA covers the group |
|---|---|---|---|---|
| Synthetic cannabinoid receptor agonists | THC (Section 1) | Potent, high-efficacy CB1 agonists | ADB-BUTINACA (2022), AFUBIATA (2023), ADB-5'Br-BINACA (2024) | Schedule II lists "synthetic cannabinoid receptor type 1 agonists" as a class, defined by core chemical structures |
| Synthetic cathinones | Amphetamine and MDMA (Lesson 6, Section 3) | Block monoamine transporters or cause monoamine release | N,N-dimethylpentylone (2022), fluoro pentedrone (2023), N-sec-butyl-pentedrone (2024) | Schedule I lists MDPV with its derivatives and analogues; Schedule III lists cathinone and methcathinone |
| Novel benzodiazepines | Diazepam and alprazolam (Lesson 4) | Enhance GABA at GABA-A receptors | Desalkylgidazepam (2022), metizolam (2023), bretazenil (2024) | Schedule IV lists "benzodiazepines, their salts and derivatives" |
The final column of Table 7.7 shows a strategy that the UNODC calls the generic approach: controlling whole groups of substances defined by chemical structure, alongside analogue laws that cover substances chemically similar to a listed drug (UNODC, n.d.). By October 2024, at least 65 countries and territories had legal responses to new psychoactive substances, and the UN Commission on Narcotic Drugs had placed 79 of them under international control (UNODC, n.d.). Canada uses group listings for all three groups. The placement matters: as of 2026, section 4 of the CDSA makes possession an offence only for substances in Schedules I to III, so possessing a novel benzodiazepine, which falls under Schedule IV, is not an offence, although selling it is (CDSA, 1996).
9.1 How Early Warning Systems Work
An early warning system tries to shorten the time between a new substance appearing in a market and the people who need to know about it (clinicians, harm reduction workers, people who use drugs and regulators) learning what it is and how dangerous it may be. Lesson 5 described the sources of information on the drug supply: seized-drug analysis, drug checking, post-mortem toxicology and wastewater testing. An early warning system links such sources together and adds procedures for assessment and response.
9.1.1 The EU Early Warning System
The European Union's system is the most formalized of the three described in this section. Regulation (EU) 2017/2101, adopted on 15 November 2017 and applied from 23 November 2018, set out the procedure: the European drugs agency, working with Europol, collects, analyzes and assesses information on new psychoactive substances from the Member States and communicates it to national focal points, Europol national units and the European Commission (Regulation (EU) 2017/2101, 2017). From 2 July 2024, Regulation (EU) 2023/1322 replaced the regulation that governed the agency and turned the European Monitoring Centre for Drugs and Drug Addiction into the European Union Drugs Agency (EUDA); it carried the early warning and risk assessment provisions over unchanged, so the procedure described here is the one in force as of 2026 (Regulation (EU) 2023/1322, 2023). The rules set deadlines for each stage, shown in Figure 7.5.
The initial report gives a first indication of the incidents in which the substance may be involved and how it is used, its chemistry and manufacture, its pharmacology and toxicology, and the involvement of criminal groups, along with any medical or industrial uses and any existing controls (Regulation (EU) 2023/1322, 2023, Article 9). The deadlines add up quickly: 5 weeks for the initial report, 2 weeks for the Commission's decision and 6 weeks for the risk assessment total 13 weeks. The EUDA's European Drug Report 2026 gives an example of the full sequence. In 2025, the EUDA's Scientific Committee carried out risk assessments on three synthetic cathinones, 2-MMC, 4-BMC and NEP, and the Commission adopted legislation placing them under EU drug control measures in January 2026 (EUDA, 2026).
9.1.2 The UNODC Early Warning Advisory
At the global level, the UNODC established its Early Warning Advisory to help countries identify and report new psychoactive substances. It serves as a repository of information from drug seizures and from toxicology casework, provides laboratory manuals, field detection kits, quality assurance and training, and supports countries in setting up national early warning mechanisms. Its data help identify the most harmful, persistent and prevalent substances, an important step in prioritizing them for review under the international drug control conventions (UNODC, n.d.). Lesson 11 explains how the World Health Organization and the Commission on Narcotic Drugs decide which substances to control internationally.
9.1.3 Canada's Arrangements
As of 2025, the same functions are shared in Canada among several organizations, summarized in the accordion below.
The Drug Analysis Service analyzes suspected illegal drugs and substances submitted by Canadian law enforcement and public health officials (Health Canada, 2024, 2025b). Health Canada publishes an annual summary of the new psychoactive substances the service identified for the first time; Interactive 7.3 presents the summaries for 2022 to 2024.
The Canadian Centre on Substance Use and Addiction coordinates CCENDU, a pan-Canadian network of community partners whose sites gather information from local partners about substance-related trends and responses (CCSA, 2025). Lesson 5 described the network's size in 2026 and drew on its June 2026 report on medetomidine; the network's newsletters, such as one on nonmedical benzodiazepines in January 2025, report new substances as sites detect them (CCSA, 2025).
Community drug checking services and coroners' toxicology testing provide information on what people are actually consuming and what is found in people who have died, as Lesson 5 showed. These sources detect substances that never reach a police seizure, and they connect a new substance to its effects on people.
9.2 Reading Canadian Data on New Substances
Health Canada's annual reports provide a direct view of what is new in Canada's drug markets. They list every substance with potential psychoactive effects that the Drug Analysis Service identified in Canada for the first time that year, by pharmacological class and by province of first identification (Health Canada, 2023b, 2024, 2025b). Use Interactive 7.3 to compare the reports for 2022, 2023 and 2024. Health Canada's report for 2025, published in August 2026 and not included in the interactive, counted 24 new substances first identified by the Drug Analysis Service, 18 of them controlled under the CDSA (Health Canada, 2026b).
Choose a year to see the substances Health Canada's Drug Analysis Service identified in Canada for the first time, grouped by class, or choose the comparison view.
Sources: Health Canada, At-a-glance: New psychoactive substances in Canada, reports for 2022 (published December 2023), 2023 (September 2024) and 2024 (August 2025). The 2022 total of 29 includes one precursor chemical and the cutting agent medetomidine; the 2023 report lists four precursors and one by-product, and the 2024 report two precursors, separately from their totals of 11.
Using Interactive 7.3, answer three questions. (a) Which class had the most first identifications in 2022, and which classes appeared in all three years? (b) The number of first identifications fell from 29 in 2022 to 11 in 2023 and 11 in 2024. Give two reasons why this fall might not mean that fewer new substances entered Canadian markets. (c) Which province accounted for most first identifications in each year, and what might that reflect?
Suggested answer: (a) Hallucinogens (8) in 2022; cannabinoids, hallucinogens, opioids, sedatives or hypnotics and stimulants all appeared every year. (b) The counts depend on which samples police and public health officials submit; the 2022 total included a precursor chemical, which the later reports list separately; the service's mandate is to report controlled substances, so uncontrolled ones may go unreported; and a substance can only be "first identified" once, so later years exclude substances already found. (c) Ontario (14, 7 and 7); it is the most populous province, and the number of samples submitted from each region shapes where substances are first found.
The reports carry caveats that apply to all monitoring data. Samples analyzed by the Drug Analysis Service may not represent the drugs circulating in Canada (Health Canada, 2023b), and the service's mandate is to report substances controlled under the CDSA, so uncontrolled substances may go unreported (Health Canada, 2024, 2025b). A first identification records when a substance was first detected in a submitted sample, which may be later than when it first appeared. The 2023 and 2024 reports each found that 7 of the 11 new substances were controlled under the CDSA (Health Canada, 2024, 2025b), and the 2022 report found 19 of 29. This is consistent with group listings at work, since a substance can be covered by law before it is first found in Canada, although the reporting mandate may also favour controlled substances. Group listings do not cover everything: the synthetic cannabinoid AFUBIATA, first identified in 2023, was one of four new substances that year that were not controlled (Health Canada, 2024).
Health Canada's report for 2024, published in August 2025, lists 11 new psychoactive substances first identified in Canada: three opioids (the nitazenes N-desethyl etonitazene and N-desethyl protonitazene, and tianeptine), two synthetic cannabinoids (ADB-5'Br-BINACA and AB-MDMSBA), two stimulants (4-methylmethylphenidate and the cathinone N-sec-butyl-pentedrone), three hallucinogens of the phenethylamine group (2C-T-2, 2C-T-7 and proscaline) and one benzodiazepine (bretazenil). Seven were first identified in Ontario, three in Quebec and one in British Columbia. Seven of the 11 were controlled under the CDSA. The most frequently identified, N-desethyl etonitazene, was found in 7 samples submitted by officials in Ontario and Prince Edward Island (Health Canada, 2025b). Use this lesson to complete three tasks.
(1) For each pharmacological class, name the established drug or drug class from Lessons 4 to 7 whose pharmacology would be the starting point for predicting its effects and risks, and the receptor involved. (2) Explain what the report can and cannot tell a public health team about whether these substances are causing harm, and name two other sources the team should consult. (3) Using Figure 7.5 as a model, describe the steps that should follow the detection of N-desethyl etonitazene in Canada, from detection to response.
(1) The synthetic cannabinoids start from THC and the CB1 receptor, but as high-efficacy agonists they may cause stronger and different effects (Sections 1 and 4). The cathinone starts from amphetamine and the monoamine transporters (Lesson 6). The phenethylamine hallucinogens start from the classic psychedelics and the 5-HT2A receptor (Section 3; Liechti, 2015). Bretazenil starts from the benzodiazepines and the GABA-A receptor (Lesson 4). The nitazenes and tianeptine are opioids, so the starting point is the mu-opioid receptor and the risk of breathing suppression (Lesson 5); 4-methylmethylphenidate relates to the prescription stimulant methylphenidate (Lesson 6).
(2) The report shows that a substance was present in samples submitted by police and public health partners. It cannot show how widely the substance is used, whether people know they are taking it, or whether it causes poisonings or deaths, and the samples may not be representative. The team should consult drug checking results, which show what people are consuming, and coroners' toxicology or poison centre data, which link substances to harm (Lesson 5).
(3) After laboratory identification, the finding is shared with partners such as CCENDU and public health agencies; an initial assessment draws together what is known about the substance's chemistry, pharmacology, harms and spread; if the risk is serious, alerts go to clinicians, harm reduction services and people who use drugs, who may need to know that a sample sold as one drug contains a potent opioid; legal control is considered last, and in this case the CDSA's Schedule I listing of benzimidazoles, their salts and derivatives, which covers the nitazenes, may already apply.
New psychoactive substances test every concept in this lesson. Their effects are predicted from the receptor pharmacology of Sections 1 and 3, their harms are judged with the same evidence that Section 2 applied to cannabis but with far less of it, and the response depends on how fast monitoring systems can turn a laboratory finding into useful warnings. The final assessment that follows draws these threads together.
Key Takeaways
- 8.1 The UNODC defines a new psychoactive substance as one not controlled by the 1961 or 1971 conventions that may pose a public health threat, and "new" means recently available rather than newly invented. Many are slight chemical modifications of controlled drugs made to get around drug laws, and they are dangerous because safety data are scarce, composition is often unknown and they are often taken with other drugs or unknowingly. They are classified by chemical structure and by effect, and the effect groups map onto the drug classes of Lessons 4 to 7.
- 8.2 Synthetic cannabinoids are potent, high-efficacy CB1 agonists that can cause agitation, psychosis and seizures; synthetic cathinones block or release monoamines as amphetamine and MDMA do; and novel benzodiazepines enhance GABA but have rarely been tested in clinical trials, so they are dangerous in combination with opioids and alcohol. As of 2026, Canada covers all three groups through group listings in the CDSA.
- 9.1 Early warning systems link laboratory detection to assessment, alerts and control. The EU rules in force as of 2026 set deadlines that add up to about 13 weeks from initial report to risk assessment, the UNODC Early Warning Advisory pools global data and supports national systems, and as of 2025 Canada relied on the Drug Analysis Service, CCENDU, drug checking and toxicology.
- 9.2 Health Canada's Drug Analysis Service identified 29 new substances for the first time in 2022 and 11 in each of 2023 and 2024, most often in Ontario. The counts depend on which samples are submitted, the service's mandate favours controlled substances, and a first identification may come after a substance has arrived, so the data show presence but not how widely a substance is used or whether it causes harm.
Reflection
In 2022, Health Canada's Drug Analysis Service identified the synthetic cathinone N,N-dimethylpentylone in Canada for the first time, along with 28 other new substances. In Europe in 2025, the cathinone N-ethylnorpentedrone (NEP) was increasingly mis-sold as another cathinone, 3-MMC, leading to unintended use and more poisonings; after risk assessments in 2025, the European Commission placed NEP and two other cathinones under EU drug control in January 2026. Explain why a person might take a new psychoactive substance without knowing it, which stages of an early warning system (detection, notification, assessment, alert and control) could reduce the harm in such a case, and one limit of relying on Drug Analysis Service data to detect such substances in Canada.
People can take a new substance unknowingly because unregulated drugs carry no reliable labels. A seller may substitute a cheaper or more available compound for the expected drug, as happened when NEP was sold as 3-MMC, or a substance may be mixed into another product. Because many new substances have been little studied, neither the seller nor the buyer may know the effects, and a cathinone with strong dopamine or serotonin effects can cause agitation, psychosis, hyperthermia or seizures.
Detection and notification matter most for this kind of harm. Laboratory identification of a new substance, shared quickly with networks such as CCENDU and public health agencies, allows an alert to reach clinicians, harm reduction services and people who use drugs, who can then know that a product may not be what it claims. Assessment of the substance's pharmacology and harms guides how urgent the warning should be, and control, as in the EU decision of January 2026, may reduce open sale but comes last and slowest.
A limit of the Drug Analysis Service data is that they come from samples submitted by police and public health partners, which may not represent what people are using, and a first identification can lag well behind first appearance. Drug checking and toxicology data are needed to show what people actually consume and whether it harms them.
Minimum 20 characters required.
Question 1: Why can synthetic cannabinoid receptor agonists produce more severe effects than cannabis?
Question 2: Under the UNODC definition, which of these is a new psychoactive substance?
Question 3: Health Canada's Drug Analysis Service identified 29 new psychoactive substances for the first time in 2022 and 11 in 2023. What is the most defensible interpretation?
Question 4: Under the EU early warning rules in force as of 2026 (Regulation (EU) 2017/2101, carried into Regulation (EU) 2023/1322), how long does the EU drugs agency have to submit an initial report on a new psychoactive substance?
Final Assessment
Bringing It All Together
This lesson examined three groups of drugs that are often discussed together but differ sharply in pharmacology and in the evidence about them. Cannabis acts on the endocannabinoid system, a retrograde braking system for synapses, through THC, a low-efficacy CB1 agonist whose concentration in products has risen several-fold since the 1970s. The 2017 NASEM review showed which claims about cannabis then rested on substantial evidence (chronic pain, nausea with chemotherapy, respiratory symptoms, crashes, low birth weight, psychosis and problem use) and which did not, later reviews have extended that picture, and the EU-GEI study linked daily use of high-potency products to the highest odds of psychosis.
Classic psychedelics, MDMA and ketamine act through serotonin 5-HT2A receptors, serotonin release and NMDA inhibition respectively. Their main risks are psychological for the psychedelics, overheating and low blood sodium for MDMA, and bladder damage with long-term non-medical use of ketamine, and the 2007 expert ratings placed their dependence potential below that of alcohol, tobacco and opioids. New psychoactive substances imitate these and other drug classes, often with little research behind them when they appear, and early warning systems in the European Union, at the UNODC and in Canada try to detect, assess and respond to them before harm spreads.
Across all three groups, the lesson has used the same questions: what does the drug do at the receptor, how strong is the evidence of harm, how does context change risk, and how well does the law match the evidence.
Key Takeaways from this lesson
- The endocannabinoid system consists of cannabinoid receptors, the endocannabinoids anandamide and 2-AG, and the enzymes that make and break them down; endocannabinoids travel backward across synapses to reduce neurotransmitter release. THC is a low-efficacy agonist at CB1 receptors that produces the high of cannabis, while CBD does not produce intoxication and may offset some, but not all, of THC's adverse effects.
- As of 2026, Canada's legal market sells products with very different THC content, from edibles limited to 10 mg of THC per package to vape cartridges, which averaged 74 percent THC in Ontario in early 2022, when legal dried flower there averaged 22 percent THC.
- THC concentrations rose by about 0.29 percentage points a year in herbal cannabis from 1970 to 2017, although seized and catalogue samples are not random; a standard THC unit of 5 mg puts products on one scale.
- The NASEM (2017) review found substantial evidence linking cannabis use to psychosis, motor vehicle crashes, respiratory symptoms, lower birth weight and problem use, and limited or insufficient evidence for many other claims. In the EU-GEI study, daily use of high-potency cannabis carried about 4.8 times the odds of psychotic disorder, and attributable fractions of up to 50 percent assume that the association is causal. Estimates of how common cannabis use disorder is depend on its definition.
- Frequent use, early starting and high-potency products are the main risk factors highlighted by the 2022 Lower-Risk Cannabis Use Guidelines, which updated the 2017 version with route-specific driving advice and new recommendations.
- Classic psychedelics act at 5-HT2A receptors, and safety reviews describe low physiological toxicity and low dependence potential, with psychological risks that depend strongly on set and setting. MDMA's serotonin release explains its risks of serotonin syndrome, hyperthermia, seizures and hyponatremia, and long-term non-medical ketamine use is associated with bladder damage.
- The 2007 expert ratings placed the dependence potential of ketamine, cannabis, LSD and ecstasy below that of heroin, cocaine, tobacco and alcohol, but the ratings are judgements by a small group in one country and year, and they did not match that country's legal classification at the time.
- New psychoactive substances are modified or repurposed compounds that imitate established drugs; synthetic cannabinoids, synthetic cathinones and novel benzodiazepines show how small chemical changes can bring large and poorly understood risks.
- Early warning systems link laboratory detection to assessment, alerts and control, and as of 2025 Canada relied on the Drug Analysis Service, CCENDU, drug checking and group listings in the CDSA; counts of first identifications show presence, not use or harm.
Core Concepts Reviewed
Section 1: the endocannabinoid system, CB1 and CB2 receptors, retrograde signalling, THC as a low-efficacy agonist, CBD, cannabis products and THC limits in Canada, rising potency and the standard THC unit.
Section 2: the NASEM weight-of-evidence categories, statistical association, the EU-GEI study, odds ratios and population attributable fractions, cannabis use disorder and the Lower-Risk Cannabis Use Guidelines.
Section 3: classic psychedelics and the 5-HT2A receptor, MDMA as an entactogen, ketamine and NMDA inhibition, acute psychological risks, ketamine-associated uropathy and expert ratings of dependence potential.
Section 4: the definition of new psychoactive substances, synthetic cannabinoids, synthetic cathinones, novel benzodiazepines, group listings, the EU Early Warning System, the UNODC Early Warning Advisory and Canadian monitoring.
The final reflection asks you to compare the kinds of evidence this lesson used and how each has shaped Canadian policy.
Reflection
This lesson used three kinds of evidence: systematic reviews and observational studies of cannabis (NASEM, 2017; Di Forti et al., 2019), expert judgement and research safety guidelines for psychedelics and dissociatives (Nutt et al., 2007; Johnson et al., 2008), and monitoring data from early warning systems for new psychoactive substances (Health Canada's annual reports; EUDA, 2026). Choose two of these three kinds of evidence. For each, describe one question it answers well, one question it cannot answer, and how it has shaped (or could shape) a real Canadian policy or guideline named in this lesson, such as the Cannabis Regulations' THC limits, the Lower-Risk Cannabis Use Guidelines or the CDSA's group listings.
Systematic reviews and observational studies of cannabis answer questions about association and its strength in real populations. NASEM's 2017 review showed which claims then rested on substantial evidence, such as the link with psychosis and with crashes, and the EU-GEI study showed a gradient by frequency and potency. These studies cannot prove causation, because confounding and reverse causation remain possible, and an attributable fraction assumes causality. Their influence is visible in the 2022 Lower-Risk Cannabis Use Guidelines, whose advice on potency, frequency, age of starting and driving maps onto this evidence, and their authors argue that regulation, such as product labelling and content rules, should support the guidelines (Fischer et al., 2022).
Monitoring data from early warning systems answer a different question: what substances are appearing, where and when. Health Canada's reports show that 29 new substances were first identified in 2022 and 11 in each of 2023 and 2024, and the EU system turned risk assessments carried out in 2025 into legal control in January 2026. These data cannot show how widely a substance is used or whether it causes harm, because samples come from police and partners and may not be representative. They support Canada's use of group listings in the CDSA, under which 7 of the 11 substances first found in 2024 were controlled at the time of the report, and they show why alerts must be combined with drug checking and toxicology to reach the people at risk.
Minimum 30 characters required.
Final Knowledge Assessment
Question 1: Which pairing of drug and main target is correct?
Question 2: How do 1 g of dried flower at 22 percent THC and a legal edible package at the 10 mg maximum compare?
Question 3: Which concept from Section 1 best explains why a synthetic cannabinoid can be more dangerous than THC even though both act on CB1 receptors?
Question 4: Which is a limitation of evidence on rising THC potency drawn from seized cannabis?
Question 5: Which comparison of psychosis risks across this lesson is accurate?
Question 6: Why might the EU-GEI attributable fraction overstate the share of psychosis caused by cannabis?
Question 7: Which 2022 Lower-Risk Cannabis Use Guidelines recommendation did its authors support with substantial evidence?
Question 8: Which statement about Canadian law as of 30 September 2026 is correct?
Question 9: Why can findings on psilocybin's safety in clinical research not be transferred directly to unsupervised use?
Question 10: Which limitation did Nutt et al. (2007) themselves note about their ratings?
Question 11: Health Canada's report for 2024 classes 2C-T-2, 2C-T-7 and proscaline as hallucinogens of the phenethylamine group. Which prediction about how they act draws correctly on this lesson?
Question 12: A person is found with a small quantity of a novel benzodiazepine such as flualprazolam. Under section 4 of the CDSA as of 2026, what is its legal position?
Question 13: How do synthetic cathinones differ in mechanism, according to Liechti (2015)?
Question 14: What is the correct order of stages in the EU Early Warning System, as set out in Regulation (EU) 2017/2101 and carried into Regulation (EU) 2023/1322?
Question 15: Health Canada's report for 2024 found that 7 of the 11 new psychoactive substances first identified in Canada were controlled under the CDSA. What does this suggest?
Glossary: Key Terms, People & Frameworks
📚 Reference page, available throughout the lesson
Definitions of the terms, frameworks and people introduced in this lesson, in the order of the sections where they first appear.