Adulterants: What Has Actually Been Found in Products

Brodifacoum: the documented case that nobody predicted human data

In the spring of 2018 emergency departments in Illinois began receiving patients with unexplained bleeding — haematuria, gum and nose bleeding, bruising, flank pain, some with intracranial or retroperitoneal haemorrhage — who had in common the recent use of synthetic-cannabinoid products. Coagulation studies showed a markedly prolonged prothrombin time correcting with vitamin K, and the responsible agent was identified as brodifacoum, a long-acting 4-hydroxycoumarin anticoagulant rodenticide, present in the products themselves and detectable in patient serum. The outbreak produced hundreds of reported cases centred on Illinois with additional cases in other states, and several deaths. The clinical problem is brodifacoum's pharmacokinetics: it is a superwarfarin with an elimination half-life measured in weeks, so treatment is not a single dose of vitamin K but high-dose oral phytonadione continued for months, with repeated coagulation monitoring and a supply and cost problem that was itself a barrier to care. Why it was in the products at all was never established with certainty; hypotheses in the literature include deliberate addition believed to prolong effects and contamination from shared equipment or supply, and both are consistent with a supply chain in which the people doing the formulation had no analytical oversight of what was in their hands. The general lesson is the important part. The contaminant was not a plausible guess. No harm-reduction list of the period included a rodenticide, no user could have detected it, and no product-identity assumption covered it. A supply chain with no analysis in it can deliver anything, and the space of possible contaminants is not limited to the things anyone thought to worry about.

Sources: Kelkar AH 2018 · Hussain N 2018 · Moritz E 2018

Insecticides on smoked material: the "bug spray" phenomenon, described as what it was human data

Reports of people applying household or agricultural insecticide to plant material and smoking it appeared in the same period and in the same populations as the synthetic-cannabinoid harms, and they are usually recounted as a punchline. They should not be. They were the furthest downstream layer of a supply chain that had already failed: people improvising with what was physically available in a corner shop, in a market where the actual product was unidentifiable, unpredictably dosed and periodically unavailable, and where no source of straight information was on offer because every institution positioned itself to withhold rather than to explain. Mockery is both cruel and analytically useless — it locates the problem in the user rather than in the information vacuum that produced the behaviour. The pharmacology of what those products contain is worth knowing plainly. Household and agricultural insecticide formulations are typically organophosphates or carbamates, which inhibit acetylcholinesterase and produce a cholinergic syndrome, or pyrethroids, which act on voltage-gated sodium channels and in high exposure produce paraesthesia, agitation, tremor and seizures; formulations also carry synergists such as piperonyl butoxide, along with petroleum distillate carriers and propellants. Inhaling combustion products of any of these is a different and worse exposure than the label warnings contemplate, because pyrolysis generates additional toxicants and because inhalation bypasses first-pass metabolism entirely. That pesticide residues do transfer into smoke rather than being destroyed by combustion has been measured directly for cannabis, with recoveries in mainstream smoke high enough to matter for several compounds, and the same physics applies to a deliberately applied insecticide at far higher loading. The correct response to this material is to explain it, which is what this page is for.

Sources: Sullivan N 2013 · Trecki J 2015 · Castaneto MS 2014

Vitamin E acetate and EVALI: a diluent chosen for the wrong property human data

The cannabis-adjacent case that makes the general point most cleanly is EVALI, the outbreak of e-cigarette or vaping product use-associated lung injury that ran through late 2019 and into 2020 and was associated principally with THC-containing vaping products from informal sources. Patients presented with dyspnoea, cough, fever, gastrointestinal symptoms and bilateral pulmonary infiltrates, many requiring intensive care. By mid-February 2020 the national case count stood at roughly 2800 hospitalised cases or deaths with 68 confirmed deaths. Investigation converged on vitamin E acetate: it was identified in bronchoalveolar-lavage fluid from 48 of 51 case patients and in none of 99 healthy comparators, a separation about as clean as this kind of epidemiology produces. Vitamin E acetate had been adopted as a cutting agent because it is viscous, oily and miscible with cannabis extract, which made diluted product look and behave like undiluted product in a cartridge — that is, it was selected for a property that defeats a consumer's visual quality check, with no consideration of its behaviour in a lung. Subsequent work showed that heating it releases ketene, a highly pulmonary-toxic gas, supplying a plausible mechanism for the injury. The structural lesson generalises well past this one compound. A diluent is not a neutral filler; it is an ingredient consumed by the same route as the active, in larger quantity than the active, and its inhalation toxicology is a separate question from its food safety. A substance can be entirely unremarkable to swallow and severely injurious to inhale, and the heated route adds degradation chemistry that no oral safety assessment covers.

Sources: Blount BC 2020 · Wu D 2020 · US Centers for Disease Control 2020* · Krishnasamy VP 2020 · Meehan-Atrash J 2021

Metals, from hardware and from catalysts human data

Two distinct metal exposures show up in this product class. The first is hardware: the metallic heating coils and associated components of vaping devices shed metal into the liquid and the aerosol, and measured aerosol concentrations of chromium, nickel, lead and manganese have been attributed specifically to the coil, with the effect varying by device and by operating power. This is a property of the device, not of the cannabinoid, and it is therefore invisible to any testing performed on the liquid before it is put into the hardware. The second is catalysis: converted-cannabinoid production uses acid catalysis, and where the catalyst is a metal species or the reaction is run in metal equipment, metal residues can carry through into the product. That matters analytically because the standard cannabis heavy-metal panel is a fixed short list — typically lead, arsenic, cadmium and mercury, the four metals that regulators were worried about for agricultural inputs — and it does not look for the metals a conversion step would plausibly contribute. A clean four-metal result on a converted product is therefore weak evidence about catalyst residues, and the analysis that would answer the question is a broader elemental scan that is rarely ordered. Surveys of retail Δ8-THC vaporiser products have reported metals alongside unlabelled compounds and processing residues, which is the expected signature of a product class formulated by processors with heterogeneous analytical capability.

Sources: Olmedo P 2018 · Meehan-Atrash J 2022 · Lin K 2026

Sold-as is not identification human data

Everything on this page reduces to one principle. A product name, a label, a brand, a seller's assurance and a community consensus are all statements about what a substance is supposed to be. None of them is a measurement. Identification means an analytical determination of what a sample actually contains, performed against an authentic reference standard of the suspected compound — the standard being the thing that turns a retention time and a spectrum into a name. This is exactly what was unavailable during the synthetic-cannabinoid era: new analogues appeared faster than reference standards could be synthesised and distributed, so even well-equipped laboratories could not put a name to material in circulation, and routine immunoassay drug screens did not detect these compounds at all. The same gap is live now in a different form: novel cannabinoids appear on labels for which no validated quantitative method and no commercially available reference standard exists, which means the number printed next to the name cannot have been measured in the ordinary sense. The practical consequences are worth stating flatly. The absence of a compound from a certificate means it was not looked for or was below the method floor, not that it is absent. An unassigned chromatographic peak is an unidentified compound present in the product. A product with no certificate is a product of unknown composition, and the only honest description of consuming it is that the identity of what is being consumed is unknown. That is not a moral judgement; it is a statement about the evidence, and it is the one piece of information that a person deciding what to do actually needs.

• Identification requires an analysis against an authentic reference standard. A label is not an analysis.
• Not detected means below the method floor for the analytes that were sought — nothing more.
• An unassigned peak is an unidentified compound being consumed.
• Routine immunoassay screens did not detect synthetic cannabinoid receptor agonists; a negative drug screen is not evidence of what was taken.
• Keep the packaging and any remaining material if someone becomes unwell — it is the only route to identification after the fact.

Sources: Castaneto MS 2015 · Trecki J 2015 · European Monitoring Centre for Drugs 2021* · Meehan-Atrash J 2022 · International Organization for Standardization / International Electrotechnical Commission 2017