25I-NBOMe

25I-NBOMe was not made to be taken. It is a laboratory tool, the N-(2-methoxybenzyl) derivative of 2C-I, synthesised by Ralf Heim at the Freie Universität Berlin to serve as a probe of the 5-HT2A receptor, then radiolabelled with carbon 11 under the name Cimbi-5 as an imaging tracer. It went from there to blotter without any stage of human evaluation having taken place, and there still is none: no controlled human administration has ever been published. Neither Shulgin nor Nichols created it, contrary to what gets said; Shulgin is the author of the 2C series it derives from, Nichols made a radioligand of it. The imbalance in the literature is the most telling fact about this molecule: it can be assayed down to the picogram in seven biological matrices, and nobody knows which dose produces which effect in a human being, nor how long it stays in the blood. This is the portrait of a substance studied through its dead.

The market, the sale on blotter like LSD, the narrow margin, the bitterness rule, the comparison with the DOx and Bromo-DragonFLY are carried by the Research chemicals page, and this entry does not redo them.

What the benzyl arm changed

The chemical gesture amounts to one addition: grafting an N-benzyl group onto a hallucinogenic phenethylamine sharply increases affinity for 5-HT2A and hallucinogenic potency. On the head-twitch response in the C57BL/6J mouse, the standard behavioural marker of hallucinogenic effect, 25I-NBOMe is 14 times more potent than 2C-I, and a selective 5-HT2A antagonist completely blocks the response.

This 14 is the top of a range and not a uniform factor, and it is the 2C-I entry that carries the full measurement. In the rat, the gap falls to 10 on wet dog shakes and to only 5 on back muscle contractions: the real in vivo range runs from 5 to 14 depending on the behaviour measured. The ratio even reverses from one measurement design to another, 25I-NBOMe being 32 times higher in binding affinity at 5-HT2A while in functional calcium potency it is 2C-I that is 8 times more potent. None of this softens the dosing trap, which remains massive, a quantity chosen for 2C-I still being a clear overdose in 25I-NBOMe. What has to be refused is the convenient single number, not the warning.

The only constants read in a primary source come from Nichols et al. 2008, who were preparing the tritiated compound: dissociation at 5-HT2A of 0.097 ± 0.02 nM in one cell line and 0.15 ± 0.01 nM in another, that is a subnanomolar affinity, roughly twenty times that of ketanserin. A more precise set of values is in circulation; it comes from a real and coherent paper, but its table could not be read and this entry does not publish it. A rarely cited detail: the radioligand labels only about 60% of the sites that ketanserin labels, which suggests a preference for a substate of the receptor rather than for the whole population. Partial or full agonist, the question is badly put, the answer depending on the functional assay used.

Two features weigh in practice. The compound has affinity for alpha-1 adrenergic receptors, constants below 300 nM, the basis of the vasoconstriction and of the stimulant component; the authors who profiled the series conclude from it that the effects could be more stimulating than those of LSD. And it has no notable affinity for 5-HT1A, where LSD and psilocin have one, an absence that serves as the common explanation for its supposedly lacking the brake of the classical psychedelics: that is a hypothesis, no source read demonstrates it. Finally, it is neither a releaser nor a reuptake inhibitor, but a direct postsynaptic agonist with no significant affinity for the serotonin, noradrenaline or dopamine transporters, which separates it radically from MDMA.

The blotter is a mixture, and its label is wrong even within the family

Three commercial blotters reputed to contain NBOMes were analysed: the majority compound differed in each one, and all of them carried traces of two or three other derivatives besides. Two adolescents convinced they had taken 25I-NBOMe had in fact taken 25C and 25B. The immediate neighbourhood is just as misleading: in the 2C family, a product advertised as 2C-I and identified on analysis as 2C-E killed one person among ten, at Blaine, Minnesota, in 2011. This count is given with its reservation, and the reservation has to be given wherever the case is told: the sources do not count the same way, one giving ten hospitalised in addition to the dead person, the other ten in all. The death itself rests, in the indexed literature, on nothing more than a table of deaths reported by the media, accompanied by its own authors' note that their accuracy cannot be verified. It is the only death in this corpus caused by a confusion of names, and that is one more reason not to harden its figure.

Ehrlich's reagent remains the most useful means, provided one understands what it says. It reacts with the indole nucleus: LSD turns purple in two to five minutes, 25I-NBOMe, a phenethylamine and not an indole, does not react. It is an asymmetric exclusion test, it says that a product is not LSD, it confirms no identity. On Marquis, 25I-NBOMe turns brown-green. A blotter carries very little active material, weak and slow reactions are the norm. The bitterness and the numbness of the mouth at the moment of taking it are the last moment at which exposure can be interrupted; the descriptor metallic, for its part, is attributed to NBOMe by one source and to LSD by another, it is of no use. Fluorescence under a UV lamp belongs to folklore, and the review that reports it says so itself.

One analytical point spills beyond the field: 2C-I is a human metabolite of 25I-NBOMe, detected in serum fifty minutes after the ingestion of a liquid that contained nothing other than 25I-NBOMe and ethanol. Finding 2C-I in someone therefore does not prove that they took any. One named method completely separates 25I-NBOH, 2C-I and 25I-NBOMe, differential pulse voltammetry on a screen-printed electrode, and that is the name to give to an analysis service.

What does the harm

The only systematic review of analytically confirmed cases covers 20 patients, all NBOMes together, 25I being the most frequent: agitation 85%, tachycardia 85%, hypertension 65%, seizures 40%, raised creatine kinase 45%, intensive care 40%, death 15%, that is three patients. This picture describes a hospitalised population, not an exposed population. The most homogeneous series, seven young men from the north-east of England confirmed in plasma, gives tachycardia 7 out of 7 and raised creatine kinase 7 out of 7, plus three seizures, three hyperpyrexias, two cases of clonus, one acute renal failure.

Rhabdomyolysis is the complication proper to this class, tied to seizures, hyperthermia and vasoconstriction, and its progression to acidosis, anuria and acute renal failure is reported as more frequent than with the other hallucinogens. One case gives the possible scale: creatine kinase at 516,455 U/L after ingestion of a dose described as relatively low.

Seizures are more frequent than with the neighbouring phenethylamines, and the comparison is internal to the family, therefore solid: in the data from the American poison centre system, single-episode seizures in 8.8% of NBOMe exposures against 3.1% for the 2C compounds, and benzodiazepines given in half the cases against a third.

The dominant documented mechanism of death is therefore the cascade of seizure, hyperthermia, rhabdomyolysis, multiple organ failure. Traumatic death during the confusional state is a second mechanism, a real one, with a fall from a balcony and autopsies that describe contusions and abrasions linked to agitation, but a minority one in what has been read: out of seven deaths on 25I-NBOMe, one is clearly traumatic and a second is disputed between the review that summarises it and the primary source. One descriptive observation deserves to be set down without being explained: in these seven deaths, the ages are 15, 15, 16, 17, 19, 21 and 23.

There is no reference lethal concentration

A 42-year-old man, drug-naive, swallowed by mistake a mouthful of a syrup filled with 25I-NBOMe. Severe agitation, coenaesthesias, complex hallucinations, complete resolution in six hours without complication. His serum on admission showed 34 ng/mL. Against that, a 17-year-old girl died at seven days with 0.25 ng/mL of whole blood, and a 15-year-old boy at three days with 0.76 ng/mL of blood drawn while he was alive. The densest review writes that even in a post mortem context the blood concentrations reported for the NBOMes remain low and are often below 0.5 ng/mL.

This overlap is read on matrices that are not strictly comparable with one another, serum against whole blood, sampling from the living against post mortem sampling, and one of the highest values comes from unpreserved post mortem aortic blood, that is from the matrix most exposed to redistribution after death. This imprecision does not weaken the conclusion, it is one more reason for it: nobody has a threshold. The outcome appears to depend far more on what happens, seizure, hyperthermia, fall, delay in reaching care, co-exposure, than on the quantity absorbed. At least one death was moreover classified as a combined intoxication with 25I-NBOMe and psilocin, which forbids using it to calibrate anything at all.

Operational corollary, and this is the sentence to keep: feeling fine, or having already taken the same quantity without harm, is not information about the risk the next time.

Dose, onset, duration: what can be said and what one must refuse to write

No validated dose range exists. What circulates comes from compilations of user reports or from retrospective reconstructions: orders of magnitude, not dosing schedules. The most cited table is inconsistent with itself for this precise molecule, its bands overlapping in two places whereas the 25C-NBOMe line, in the same table, does not overlap. The honest formulation is therefore a single order of magnitude, from a few tens to about a thousand micrograms by the sublingual route, with no steps.

The preferred routes are sublingual, buccal and nasal, because of an extensive first-pass metabolism. It has to be said correctly, because the usual formulation is dangerous: oral bioavailability is reduced, it is not nil, and the oral route is not harmless. The case of the swallowed syrup is a clear-cut poisoning, and ingestion appears as a route in fatal cases, including that of the blotter swallowed by the 17-year-old girl. One case does not measure a bioavailability, and the high concentration may be down to the strength of the liquid, 2800 µg/mL, rather than to good absorption. But believing in oral inactivity is in itself a risk factor, since that belief motivates a redose.

Onset is the central trap. The compilations give 45 to 90 minutes sublingually, the swallowed case reports the first signs at 30 minutes, and the other cases allow no reliable reconstruction. A long onset produces the mechanism already documented for the DOx: the user concludes that nothing is coming and takes more. The compiled duration is 5 to 10 hours, residual effects 6 to 24 hours. In poisoning, the duration is another matter: six hours in the naive case, seven hours for an isolated seizure, forty-eight hours for the vital signs to return to normal, with persistent aggression.

There exists no human pharmacokinetic data, no published half-life, no measured bioavailability, no concentration-time curve, and that is a result in itself. The only tolerance data comes from the rat, where it is not uniform and touches the sought-after effect without touching the others; if that transposed, tolerance would encourage redosing on a molecule whose toxicity, for its part, does not diminish.

Interactions

CYP3A4 is the best-founded interaction, and it has never been observed. Two independent teams have characterised the metabolism and each concludes there is a risk of interaction; CYP3A4 is the only point on which all the sources converge, and one writes explicitly that users may be exposed to an interaction if they take it together with a potent CYP3A4 inhibitor, a category that includes common medicines, certain azole antifungals, certain macrolides, certain antiretrovirals. The direction of the effect is predicted, slowing elimination increases exposure at unchanged quantity, on a product dosed in micrograms and with a narrow margin. Its size is quantified nowhere, and there is no documented human case.

From this follows an inference that exists as yet in none of our texts and that must be given for what it is. The beta-carbolines of Syrian rue inhibit CYP3A4 in vitro, and CYP3A4 is the major enzyme for 25I-NBOMe. Two sources read, neither measurement nor case: it is a piece of reasoning, not a fact, and it is aimed at a narrow margin.

Lithium. This is the only death in the whole literature consulted that its authors attribute to a drug interaction. A 17-year-old girl presents in status epilepticus shortly after ingesting a blotter at a concert, develops hyperthermia, acidosis, rhabdomyolysis, acute renal failure and then cerebral oedema, and dies at seven days. Whole blood: 25I-NBOMe 0.25 ng/mL, lithium 0.34 mmol/L, that is below the usual therapeutic range. Lithium lowers the seizure threshold. Causality is not demonstrated, and the real confounder is that the picture observed is already the one the molecule produces on its own. The full paper for this case exists and could not be read; what is known of it comes from the table of a review, which incidentally dates the case to 1988, an obvious typographical error since the molecule only appears on the market in 2012.

SSRIs, MAOIs, serotonin syndrome. Two statements to be kept separate. No human case of interaction between an NBOMe and an SSRI or an MAOI has been found, and the closest joint exposure the literature offers, a driver declaring an unnamed antidepressant and in whom only 25C-NBOMe was found, reports no serotonin syndrome. But serotonin syndrome under 25I-NBOMe alone is described: a dedicated systematic review names it among the substances involved, with the sublingual route on blotter. We have read only its abstract, the list of cases and of their concomitant medications remaining inaccessible; that is the hole that persists on this question. Caution towards serotonergics remains justified, not by reported cases that do not exist, but because the picture of 25I-NBOMe, hyperthermia, clonus, rigidity, seizures, is already that of a serotonergic toxicity, which would make any addition hard to distinguish and to treat.

The seizure threshold, added up. The seizure signal is the strongest of the lot here, and yet it is a shared signal: of the fourteen substances in this corpus, ten carry one, six on counted human cases, this one included, and four on thinner support, an animal signal, an isolated case or an unquantified clinical mention. Syrian rue is one of those four, and therefore falls twice in this entry, once through CYP3A4 and once through the threshold. Crossing 25I-NBOMe with one of the nine others very often amounts to lowering twice the threshold that here commands the fatal cascade.

Stimulants. The alpha-1 affinity is the basis of the vasoconstrictor component. Adding a sympathomimetic stacks on the same axis. No case of this precise combination is documented, the reasoning is direct.

Co-exposure is the rule and that is a result in itself. Among the deaths read there is THC in several cases, psilocin in one, methamphetamine with two other NBOMes in another, lithium in the last. In practice, this molecule is rarely taken alone, which makes any causal attribution difficult.

What remains afterwards

A retrospective analysis of 58 reports collected by a Polish pharmacovigilance centre between 2013 and 2020 finds 15 cases, in patients aged 19 to 26, where symptoms persisted long after stopping: pseudohallucinations, bizarre delusions, derealisation, and sometimes the appearance or the worsening of a depression. Reported duration from two months to two years, more frequent in regular users, pharmacological treatment necessary in some. Risk factors noted: severe visual pseudohallucinations, delusions, derealisation or depersonalisation occurring immediately after taking it.

The nature of the object, not to be over-interpreted: a retrospective pharmacovigilance series, without controls, without an exposure denominator, with the reporting bias proper to this kind of collection. It is nonetheless the only quantified data that exists on the prolonged effects of this molecule. In the rat, chronic treatment produces a brain accumulation, DNA damage at 72 hours and a fall in glial, non-neuronal, cells in the frontal cortex; none of this has been transposed to humans.

Harm reduction

Contraindications deduced from established mechanisms, and it has to be put that way. Epilepsy or a treatment that lowers the seizure threshold, that is the most solid one. Treatment with lithium. Treatment with a potent CYP3A4 inhibitor. Heart disease or hypertension. Renal risk, rhabdomyolysis being the signature complication and renal failure its outcome. A context of heat and prolonged exertion, which combines with the documented hyperthermia. A history of psychosis, or a previous episode of persistent perception disorder.

Warning signs, from the earliest to the most serious.

• A marked bitter taste or numbness of the tongue and mouth at the moment of taking it, the only moment at which exposure can still be interrupted.
• Agitation that climbs instead of settling, confusion, aggression, disorganised behaviour.
• Persistent tachycardia and hypertension.
• Cold, pale, painful extremities, a sign of vasoconstriction.
• Rising temperature, heavy sweating.
• Clonus and muscular rigidity, which point towards serotonergic toxicity. Hyperreflexia, often cited, is rare with this class and must not be used as a marker.
• A seizure, even a single brief one: a medical emergency, it is the documented point of entry into the cascade.
• Dark or red urine, intense muscle pain, a fall in urine volume: rhabdomyolysis already established.

A sober third party, for two reasons and not one. They prevent falls and trauma during the confusional state, and above all they make it possible to spot a seizure and call early, which bears on the most frequent mechanism of death.

In the emergency department, saying what was taken has real clinical value. There is no antidote, all treatment is symptomatic: intravenous benzodiazepines for sedation and seizures, mechanical cooling, early fluid loading aimed at a urine output of 200 to 300 mL/h in case of rhabdomyolysis. And the laboratory will not find the product on its own: rapid immunoassays do not detect it reliably, gas chromatography without derivatisation is inadequate, concentrations are of the order of 0.1 to a few ng/mL and detection has to target the metabolites rather than the parent compound. A negative screen excludes nothing. Keeping the bag or the blotter materially changes the management.

Legal status and where things stand

In France, the order of 6 November 2015 names 25I-NBOMe individually in annex III of the list of narcotics and removes it from annex IV: this is a reclassification, it was already listed there. The same text sets out in annex IV a generic structural definition covering N-benzylated phenethylamines, but it excludes by name 25I, 25B and 25C, precisely because they are henceforth named one by one. In the United States, temporary placement in Schedule I on 15 November 2013, extended in 2015. Internationally, Schedule I of the 1971 Convention, with review dates that diverge according to our sources, 2014 or 2016, without our having been able to settle it, the primary documents being in PDF.

One last piece of honesty: we found no recent prevalence data. The clinical literature is massively concentrated on 2013 to 2018. This entry therefore says neither that the phenomenon is current, nor that it is over.

Sources

• J.B. Zawilska, M. Kacela, P. Adamowicz (2020), NBOMes, highly potent and toxic alternatives of LSD, Frontiers in Neuroscience (full text; the densest source: dose table, deaths table, management, detection. Contains a date typo, 1988, for the lithium case).
• D.E. Nichols et al. (2008), tritiated high-affinity radioligand for 5-HT2A, Bioorganic & Medicinal Chemistry (Kd of 0.097 and 0.15 nM, 60% of ketanserin's sites).
• A. Rickli et al. (2015), receptor interaction profiles of NBOMe derivatives, Neuropharmacology. Abstract only read, tables not consulted: the per-compound constants that circulate come from this paper and are not reproduced here.
• A.L. Halberstadt (2017), pharmacology and toxicology of N-benzylphenethylamines, Current Topics in Behavioral Neurosciences (review of 51 cases; rhabdomyolysis).
• A.L. Halberstadt and M.A. Geyer (2013), head-twitch response, Neuropharmacology (factor of 14 in the C57BL/6J mouse).
• L.M. Nielsen et al. (2017), cytochrome P450 enzymes involved in the metabolism of 25I-NBOMe and 25I-NBOH, Drug Testing and Analysis (source of the CYP3A4 warning).
• A.T. Caspar et al. (2015), metabolism and toxicological detection, Analytical and Bioanalytical Chemistry (68 metabolites).
• J. Suzuki et al. (2015), toxicities associated with NBOMe ingestion, a review of the literature, Psychosomatics (20 patients, all the percentages).
• S.L. Hill et al. (2013), analytically confirmed severe clinical toxicity, case series, Clinical Toxicology (seven patients, raised creatine kinase in all seven).
• M. Hermanns-Clausen et al. (2017), bad trip on 25I-NBOMe, SPICE II Plus project, Clinical Toxicology (the decisive case: naive subject, 34 ng/mL, resolution in six hours).
• W. Waldman et al. (2018), severe toxicity confirmed by LC-MS-MS, Acta Biochimica Polonica (creatine kinase at 516,455 U/L).
• L.M. Lowe, B.L. Peterson, F.J. Couper (2015), first analytically confirmed death in Washington State, Journal of Analytical Toxicology (death classified as a combined intoxication with psilocin).
• V.B. Kueppers and C.T. Cooke (2015), 25I-NBOMe related death in Australia, Forensic Science International (28 µg/L, unpreserved post mortem aortic blood).
• J.P. Walterscheid et al. (2014), pathological findings in two deaths, American Journal of Forensic Medicine and Pathology.
• K.G. Shanks, T. Sozio, G.S. Behonick (2015), fatal intoxications with 25B and 25I-NBOMe in Indiana, Journal of Analytical Toxicology.
• D. Schetz, A. Schetz, I. Kocić (2022), retrospective analysis of the Neverending Trip, Biomedicine & Pharmacotherapy (58 reports, 15 persistent cases).
• L.K. Laskowski et al. (2015), 25C and 25B sold as 25I, Journal of Medical Toxicology. The creatine kinase peak and the length of hospital stay in this paper concern a case of 25C-NBOMe and must never be entered into the 25I file.
• J.L. Poklis et al. (2015), analysis of NBOMe derivatives on blotter paper, Journal of Analytical Toxicology (the majority compound differs from one blotter to another).
• S.J. Stellpflug et al. (2014), case report with confirmation, Journal of Medical Toxicology (rapid immunoassays do not trigger a reliable positive result).
• F. Schifano et al. (2021), novel psychoactive substances and the occurrence of serotonin syndrome, systematic review, Experimental Neurology. Abstract only read: the review names 25I-NBOMe, the list of cases and of their concomitant medications could not be read.
• A. Zwartsen et al. (2017), measurement of monoamine transporter inhibition by NPS, Toxicology in Vitro. Abstract only read, numerical values not consulted.
• D. Luethi and M.E. Liechti (2020), designer drugs, mechanisms of action and adverse effects, Archives of Toxicology (partial or full agonists depending on the functional assay).
• A. Ettrup et al. (2010), radiosynthesis and evaluation of 11C-CIMBI-5, Journal of Nuclear Medicine (tracer evaluated in the rat and the pig; the tracer that went to humans is Cimbi-36, which is 25B-NBOMe).
• D.E. Nichols and C.S. Grob (2018), is LSD toxic?, Forensic Science International (a death attributed to LSD was in all likelihood due to a substituted substance, the authors citing 25I-NBOMe).
• J.W. Rajotte, J.-P.F.P. Palmentier, H.R. Wallage (2017), driver under 25C-NBOMe, Journal of Forensic Sciences (antidepressant declared, unnamed and not found, no serotonin syndrome).
• DanceSafe, Always Check Your LSD (Ehrlich, Marquis, taste; harm reduction source, not academic).
• Order of 6 November 2015 amending the order of 22 February 1990, Légifrance (articles 2, 3 and 4).
• Erowid, NBOMe Series Vault (corpus of user reports, to be read on site; its licence forbids reproduction. Page returning HTTP 403 when we visited, not read).
• Y. Umemura, T. Andrew, V. Jacobs, A. Giustini, L. Lewis, J. Filiano (2015), Fatal 25I-NBOMe Intoxication: A New Recreational Risk, Academic Forensic Pathology, DOI 10.23907/2015.009. Text not read: journal not indexed in PubMed, publisher's page unreadable. This is the lithium case, and its detail comes to us from Zawilska's table and from secondary abstracts. The corresponding conference abstract, Neurology 82(10 Suppl), P1.342, 2014, returns HTTP 403, not read.
• A.F. Belchior de Andrade and J. Gonzalez-Rodriguez (2019), differential pulse voltammetry on a screen-printed electrode, The Analyst 144(9):2965-2972, PMID 30839986. Not read: the method that distinguishes 25I-NBOH, 2C-I and 25I-NBOMe is cited for its existence, not for its results.
• M. Dean et al. (2013), on 2C series poisonings, source of the product advertised as 2C-I and identified as 2C-E at Blaine, Minnesota. Full text not read, known through a second-hand reading in our own corpus.
• WHO (Expert Committee on Drug Dependence) and EMCDDA reports on 25I-NBOMe: not read, available in PDF only.