Why this poison is interesting
Every other page in this library names a poison. Paracetamol, digoxin, ethylene glycol — the substance is known, its label or its literature can be consulted, and its kinetics can be quoted. This page cannot do any of that, and the reason is the entire point of it. Ford, Tai and Fantegrossi describe synthetic cannabinoids as a highly structural diverse group of compounds, easily synthesized, whose active ingredients were determined to act via CB1 cannabinoid receptors, similar to those of Δ9-tetrahydrocannabinol, and which are often abused ... to elude detection in drug tests due to their lack of structural similarity to Δ9-THC.1
That last clause matters more than it looks. They are abused to elude drug tests because they lack structural similarity to THC1 — a statement about why users choose them, not about what chemists intended. Whatever the intent, the consequence is the same: the property that makes them undetectable is precisely what removes any assurance that they behave like THC. The name 'synthetic cannabinoid' encourages a false inference: that they are cannabis, made in a laboratory. They are compounds that happen to reach the same receptor, from an entirely different chemistry, with entirely different efficacy.
The second reason is pharmacological and is the single most useful thing on this page. THC is a partial agonist at the CB1 receptor. Many of the synthetic compounds are full agonists. A partial agonist has a ceiling: however much is present, the receptor response cannot exceed a fraction of its maximum. A full agonist has no such ceiling. This is the same argument the amphetamine page makes about release versus reuptake blockade, and the SSRI page makes about MAO inhibition — the dangerous property is the one without a ceiling, and it recurs so often in this library that it has become a theme.
A poison is a drug whose kinetics have escaped its pharmacology.
The toxic principle
The CB1 receptor is a G-protein-coupled receptor densely expressed in the brain — in the basal ganglia, cerebellum, hippocampus and cortex — where it sits presynaptically and suppresses the release of other transmitters. That anatomy accounts for cannabis's effects on movement, memory and perception, and it accounts for the synthetic compounds' effects too. What differs is how hard the receptor is pushed.
Three further properties of the class are worth stating because each has a clinical consequence, and none requires knowing the specific compound.
CB1 receptors are sparsely expressed in the brainstem cardiorespiratory centres compared with the cortex, hippocampus, basal ganglia and cerebellum, and that is the usual explanation for cannabis not causing respiratory arrest. It is worth noting that the synthetic compounds do not change this — they act at the same receptor — so respiratory depression in a synthetic cannabinoid presentation should raise the question of a co-ingestant, particularly an opioid, rather than being attributed to the cannabinoid.
Toxicokinetics
This is the shortest and least satisfying kinetics table in the library, and the fourth column explains why in every row. It is included rather than omitted because the absence of kinetic information is itself a clinical fact with consequences.
| Parameter | Therapeutic | In overdose | Why it changes |
|---|---|---|---|
| Which compound | — | Unknown. A highly structural diverse group of compounds, easily synthesized1 | Everything below follows from this row. No half-life, volume of distribution or clearance can be quoted for a substance that has not been identified, and identification requires a laboratory the patient will never see in time. |
| Absorption | — | Smoked in almost all cases; effect within minutes | The one row that is reliable. Pulmonary absorption is near-instantaneous and route-limited rather than compound-limited, so onset is fast whichever compound it is. A patient who deteriorates an hour later has redosed or taken something else. |
| Potency | — | Varies by orders of magnitude across the class | This is the row that makes 'how much did you take?' meaningless. The same visible quantity of the same-looking product can deliver very different doses, and users cannot titrate against an unknown. |
| Efficacy at CB1 | THC is a partial agonist2 | Many synthetic compounds are full agonists2 | The pharmacological difference that defines the class, and the reason a receptor that limits cannabis toxicity does not limit this. A ceiling that belongs to the ligand and not to the receptor disappears when the ligand changes. |
| Metabolism and metabolites | — | Active metabolites are common in this class | Extends the effect beyond the parent's presence and is one plausible reason presentations can be prolonged. Compound-specific, therefore unknowable at the bedside. |
| Detection | — | Not detected by routine cannabis immunoassays — these compounds lack structural similarity to THC, which is why they are abused to elude drug tests1 | A negative drug screen is not evidence against this diagnosis; it is the reason these compounds are chosen. Confirmatory identification needs mass spectrometry, is not available acutely, and is always retrospective. |
| Dialysability | — | Never assessed. EXTRIP has published no recommendation covering synthetic cannabinoids3 | The absence is uninformative and untestable: you cannot assess the dialysability of a class whose members change every year. These are lipophilic compounds and would be poor candidates in any case. |
Metabolism and the metabolites
No specific metabolic pathway can be drawn, because the pathway depends on the compound. What can be drawn is the structure of the problem, and it is worth drawing because it is the structure that makes this class dangerous.
- An unidentified compound, sprayed onto plant materialOne of hundreds. A highly structural diverse group of compounds, easily synthesized1. The distribution on the plant material is itself uneven
- Systemic circulationThe one reliable kinetic fact on this page
- Cannabinoid effect without a ceilingPsychotropic actions via CB1 cannabinoid receptors, similar to those of Δ9-THC1 — similar in kind, unlimited in degree
- Hepatic metabolism, compound-specificActive metabolitesCommon in this class, and they prolong the effect beyond the parent
This class is also concentrated in specific populations, and that concentration is not incidental. Synthetic cannabinoids are used disproportionately in prisons and among people who are homeless, for reasons that follow directly from the pharmacology described above: they were designed to evade drug testing, they are cheap, and small quantities are easily concealed. The result is a poisoning whose typical patient has poor access to healthcare, is unlikely to know what they have taken, and is often unable to give a history at all — which compounds every diagnostic difficulty on this page.
Elimination and accumulation
Nothing quantitative can be said, and this section states what can be inferred structurally instead — flagged throughout as inference rather than dressed up as pharmacokinetics.
Withdrawal is described and is a genuine clinical entity — irritability, insomnia, sweating, nausea and craving in dependent users, more marked than cannabis withdrawal. It is consistent with full agonism producing greater receptor adaptation than a partial agonist can. It is not, however, the medical emergency that GHB withdrawal is, and this page does not overstate it: the comparison is included because the two drugs share a population and a setting, not because the syndromes are equivalent.
Target organs — and why those
This organ list is shorter and more heavily hedged than any other in the library, and that is the finding. A card in this section is supposed to answer why this organ. For most of these it cannot, and saying so is more useful than manufacturing an answer.
Brain — CB1 receptors
TargetPresynaptic CB1 receptors in cortex, hippocampus, basal ganglia and cerebellum
Why hereThe only card on this page whose 'why' is properly answerable. CB1 receptors sit presynaptically and suppress transmitter release; the compounds are agonists at them, producing psychotropic actions via CB1 cannabinoid receptors, similar to those of Δ9-THC.1 Full rather than partial agonism removes the ceiling that limits cannabis, which accounts for the difference in severity if not for the specific syndromes. Established
At the bedsideSevere agitation, psychosis, dissociation, catatonia-like states, and the profound unresponsiveness that gives the class its street reputation. Recovery is usually complete.
Brain — seizure threshold
TargetNot established
Why hereSeizures are the single most consistent serious finding in synthetic cannabinoid presentations, and cannabis does not cause them. That contrast is strong evidence the effect is real and related to the difference in agonism. The mechanism is not established — the cited review states that the dangerous adverse effects of these compounds occur by as of yet unknown mechanisms.1 The convulsant effect of two members of the class has been attributed to CB1 agonism in mice2 — attenuated by rimonabant, by THC, and by CB1-downregulating regimens — which is a model reproducing the syndrome rather than a bare absence, so this card is marked as inference rather than as undemonstrated teaching. The human mechanism remains unestablished. Inferred
At the bedsideGeneralised seizures, sometimes recurrent. A seizure in a cannabis-like presentation should raise synthetic cannabinoids specifically, because it argues against cannabis and towards this class.
Cardiovascular system
TargetNot established
Why hereTachycardia and hypertension are near-universal and are consistent with what cannabis does at lower intensity. Reported myocardial infarction and arrhythmia in young users are a different matter and have no established mechanism in this class. Adulterants and co-ingestants cannot be excluded in any individual case, which is a genuine limitation rather than a hedge. Inferred
At the bedsideTachycardia, hypertension, chest pain. ECG is reasonable; attributing an abnormality to the cannabinoid specifically is not usually possible.
Kidney
TargetNot established
Why hereA card recording an epidemiological signal, not a mechanism. Clusters of acute kidney injury tied to particular batches have been reported, which is the characteristic way harms from this class are discovered. Whether the injury is direct, rhabdomyolytic from agitation and seizure, or caused by something else in the product is not established. Inferred
At the bedsideCheck renal function and creatine kinase. The clustered, batch-associated pattern is itself diagnostically useful — several presentations from one locality within days should prompt a public health conversation as much as a clinical one.
Gastrointestinal tract
TargetCB1 receptors centrally and in the gut
Why hereSevere vomiting is common, and cannabinoid hyperemesis is well described with cannabis itself — a paradox, given that cannabinoids are antiemetic at lower doses and by different routes. The paradoxical hyperemesis mechanism is unresolved even for cannabis, where it has been studied far more, so it cannot be considered established here. Inferred
At the bedsideProtracted vomiting, sometimes with abdominal pain. Dehydration and electrolyte disturbance follow, and contribute to the renal picture above.
Timeline of effects
- Seconds to minutesOnsetWhat you seeRapid alteration in behaviour and consciousness — from profound unresponsiveness to severe agitation.What is happeningSmoked, so absorption is pulmonary and near-instantaneous. Full agonism at CB1 without the ceiling that limits THC.
- 5–60 minPeakWhat you seeAgitation, psychosis, dissociation, vomiting, tachycardia, hypertension. Seizures may occur here — and cannabis does not do this.What is happeningVery dangerous adverse effects occurring by, as of yet, unknown mechanisms.1 The page cannot say more than the review does.
- 1–6 hResolution — usuallyWhat you seeMost presentations settle with observation and sedation.What is happeningRedistribution and metabolism of an unidentified compound. No half-life can be quoted, and the honest position is that duration is unpredictable.
- Hours, sometimesProlonged effectWhat you seeA minority remain agitated or unwell far longer than expected.What is happeningActive metabolites are common in this class and are the usual explanation, though not one that can be confirmed for a compound nobody has identified. Compound potency varying by orders of magnitude is an equally plausible account.
- Same batch, next portionThe unpredictabilityWhat you seeA user poisoned by the portion after several uneventful ones.What is happeningUneven distribution of a powder sprayed onto plant material. The dose varies within a single package. This is not a phase of a poisoning so much as a property of the product.
- DaysWithdrawal in dependent usersWhat you seeIrritability, insomnia, sweating, nausea, craving.What is happeningReceptor adaptation to sustained full agonism. More marked than cannabis withdrawal, and not the emergency that GHB withdrawal is.
What the mechanism predicts at the bedside
- Do not model this as strong cannabis. THC is a partial agonist and many of these compounds are full agonists — a difference in kind, not degree.
- A seizure in a cannabis-like presentation argues for this class, because cannabis does not cause them.
- A negative cannabis screen means nothing. These compounds lack structural similarity to THC, which is why they are abused to elude drug tests.1
- Respiratory depression should raise a co-ingestant, usually an opioid. CB1 receptors are not present in the brainstem respiratory centres, and that does not change with a synthetic agonist.
- Dose history is uninformative. Potency varies by orders of magnitude between compounds and the drug is unevenly distributed within a single package.
- Ask where the product came from and whether others have been affected. Batch-associated clusters are how the harms of this class are usually identified, and a second similar patient is a public health signal.
- Check renal function and creatine kinase, given the reported acute kidney injury clusters and the agitation.
- Benzodiazepines for agitation, and conventionally for seizures — but note that in the one mechanistic study cited here, diazepam attenuated pentylenetetrazol convulsions and did not alter those induced by two synthetic cannabinoids, and the authors conclude benzodiazepines may not be effective treatments.2 That is a mouse finding, and it is the only mechanistic evidence there is.
- Expect a short course, and be prepared to be wrong. A minority are prolonged, and no kinetic figure exists to predict which.
- Consider the setting. Use is concentrated in prisons and among people experiencing homelessness, and the history may be unobtainable for reasons that have nothing to do with the drug's effects.
The antidote, from the poison's side
There is no antidote, and unlike most pages in this library there is not even a treatment aimed at the receptor. Management is sedation and observation, which is the same answer the GHB and ketamine pages reach, and here it is reached for a worse reason: not because the poisoning is self-limiting and benign, but because nobody knows enough to do better.
- Benzodiazepines
- For agitation, and conventionally for seizures. As on the cocaine and amphetamine pages, they reduce agitation, muscle activity and therefore heat production together. For the seizures specifically the evidence is worse than that reassuring sentence suggests: in mice, diazepam attenuated pentylenetetrazol convulsions but did not alter those produced by JWH-018 or 5F-AB-PINACA, and the authors conclude that benzodiazepines may not be effective treatments.2 An animal finding, and the only mechanistic evidence in existence — reported here rather than omitted, because omitting it is how a page ends up more confident than its own sources.
- Supportive care and observation
- The treatment. Airway, fluids, temperature, and a period of observation whose length cannot be justified from kinetics because there are none.
- A CB1 antagonist
- Chemically possible and clinically unavailable. Rimonabant, a CB1 antagonist/inverse agonist, was licensed as an anti-obesity drug and withdrawn for severe psychiatric adverse effects including depression and suicidality. It has not been developed or used as an antidote, and its withdrawal history is the reason it is unlikely to be — the same shape as the clonidine page's central alpha-2 antagonist and the antipsychotic page's D2 agonist: an antidote that would probably work and that nobody wants to use. Rimonabant and a central alpha-2 antagonist are alike in being effectively unavailable; the D2 agonist is available and merely a bad idea. Naming the three is more useful than counting them, since the count is what drifts.
- Naloxone, flumazenil
- No role. Named here only because an unresponsive patient of unknown history will usually have received both, and their failure to work is not diagnostically informative.
- Activated charcoal
- No role. The route is almost always inhalation.
- Extracorporeal removal
- Never assessed and not assessable.3 These are lipophilic compounds and the class changes composition faster than any evidence base could be built.
Critical appraisal
- The 'unknown mechanism' badge is unusual in this library and is the strongest kind of citation for the doubt available. Ford, Tai and Fantegrossi state that these compounds produce very dangerous adverse effects occurring by, as of yet, unknown mechanisms.1 It is applied to the downstream clinical claims — the seizures above all — and not to the CB1 agonism, which the same review establishes. A reader should notice that the badge here marks an absence the authors themselves declare, rather than one this page has inferred.
- The partial-versus-full agonist distinction is the load-bearing claim on this page, and at audit it acquired the direct citation an earlier draft said it lacked. Wilson and colleagues name partial CB1R agonist ∆9-tetrahydrocannabinol and full CB1R agonist SCBs JWH-018 and 5F-AB-PINACA in one sentence and demonstrate that the efficacy difference produces a qualitative behavioural difference — THC did not elicit convulsions in mice, both synthetic compounds did.2 The pharmacological classification is therefore cited; the human consequence remains inference. The earlier draft apologised at length for a claim its own reference list already supported, which is its own kind of error.
- No compound-specific information appears anywhere on this page. That is deliberate. Naming JWH-018 or AB-FUBINACA and attaching kinetics to it would give a false impression that the identity is knowable at the bedside, which is the misconception this page exists to correct. The compounds appear in the search terms only.
- The 'active metabolites are common' claim is class-level and is not supported by a citation here. It is widely reported for several members of the class and is offered as inference — including in the elimination section, which explicitly flags that it sits awkwardly with the commonly reported short duration.
- The acute kidney injury clusters are described as an association without a mechanism, which is exactly what they are. No frequency, no batch and no compound is named, because those figures are jurisdiction- and year-specific and would be quoted as though general.
- The prison and homelessness epidemiology is stated without a citation and is included because it changes clinical practice — history-taking, follow-up and safeguarding — rather than because it is a pharmacological claim. It should be read as clinical context, not as data.
- The rimonabant paragraph describes a real licensing history and is used to make a pharmacological point about why the obvious antidote does not exist. Nothing on this page suggests it should be tried, and its psychiatric adverse effect profile is the reason.
- No dose, purity, potency multiple or lethality figure appears on this page. For a class of hundreds of compounds with unknown content and uneven distribution within a package, any such number would be actively misleading as well as out of scope.
- EXTRIP has addressed nothing in this class3, and — unlike every other page in this library — the absence is not merely uninformative but permanent: a class whose members are replaced yearly cannot accumulate the evidence base an EXTRIP review requires.
References
- 1Ford BM, Tai S, Fantegrossi WE, et al. Synthetic pot: not your grandfather's marijuana. Trends in Pharmacological Sciences 2017 Mar;38(3):257–76. PMID 28162792. doi:10.1016/j.tips.2016.12.003. (Describes synthetic cannabinoids as producing psychotropic actions via CB1 cannabinoid receptors similar to those of Δ9-THC; often abused to elude detection in drug tests due to their lack of structural similarity to Δ9-THC; a highly structural diverse group of compounds, easily synthesized, which produce very dangerous adverse effects occurring by, as of yet, unknown mechanisms; falsely marketed as safe marijuana substitutes, and available evidence indicates K2/Spice products are clearly not safe marijuana alternatives.)
- 2Wilson CD, Tai S, Ewing L, et al. Convulsant effects of abused synthetic cannabinoids JWH-018 and 5F-AB-PINACA are mediated by agonist actions at CB1 receptors in mice. Journal of Pharmacology and Experimental Therapeutics 2019 Feb;368(2):146–56. PMID 30420360. doi:10.1124/jpet.118.251157. (An animal study, cited here only to record that the convulsant effect has been attributed to CB1 agonism in mice — it is not a human demonstration, and the seizure mechanism in humans remains unestablished.)
- 3EXTRIP Workgroup — published recommendations index. Extracorporeal Treatments in Poisoning Workgroup. Cited for the absence of any recommendation covering synthetic cannabinoids. extrip-workgroup.org/recommendations