2C-I

The dominant danger of 2C-I is not in its pharmacology, it is in its name: two confusions of language and two analytical mechanisms attach the 2C-I label to something else, and a fifth confusion, purely commercial, has killed. In this batch of fourteen substances, the one that has cost a life is 2C-E sold as 2C-I. The molecule, when it really is the molecule, is a psychedelic phenethylamine of Shulgin's 2C series, a 5-HT2A receptor agonist, and its literature is so thin that this void is the second result of this page: nothing in human pharmacology, nothing in pharmacokinetics, in imaging or in tolerance, and not one death attributed to it. There are four case reports in which 2C-I has been detected in a human, and none is an analytically established exposure to 2C-I alone.

The name, and what it covers

25I-NBOMe is the N-benzyl derivative of 2C-I, 2-(4-iodo-2,5-dimethoxyphenyl)ethanamine, and it is also called 2C-I-NBOMe, including in the titles of published scientific papers: the two names differ only by a suffix. The street names Smiles and N-bomb designate 25I-NBOMe, never 2C-I, and the two North Dakota deaths of June 2012, often told as 2C-I deaths, are 25I-NBOMe deaths.

The potency gap between the two depends on what is measured, and the right number is worth more than a convenient order of magnitude. In vivo in the rodent, 25I-NBOMe is 5 to 14 times more potent: 14 on head twitches in the mouse, 10 on wet dog shakes in the rat, only 5 on back muscle contractions. It is 32 times higher in 5-HT2A binding affinity, but in calcium functional potency the ratio reverses and it is 2C-I that is 8 times more potent. A dose chosen for 2C-I applied to 25I-NBOMe remains a massive overdose, it is not a uniform factor of ten.

And 25I-NBOMe is active by the oral route: first-pass metabolism reduces its bioavailability, and what remains is enough. A 42-year-old man drank a mouthful of a syrup containing only 25I-NBOMe and ethanol, and had a severe poisoning, serum at 34 ng/mL; the two North Dakota deaths were ingestions. The sentence in circulation, according to which a swallowed NBOMe does nothing, comes from an element its own authors describe as anecdotal, and it can kill whoever believes it.

Then come two analytical mechanisms, and they are the least known. On seized product: 25I-NBOH, another derivative sold on blotter, is thermolabile and degrades in the injector of the gas chromatograph, giving a majority peak read as 2C-I; none of the settings tested by this team, temperature, split, flow, prevents it, and only derivatisation avoids it. A GC-MS result displaying 2C-I does not prove that the sample contained 2C-I. On biological fluid, whatever the method: fifty minutes after that mouthful of syrup, the patient's serum contained 25I-NBOMe at 34 ng/mL and 2C-I at 12 ng/mL. 2C-I is an in vivo human metabolite of 25I-NBOMe, and an impeccable LC-MS/MS will therefore find some in someone who has never taken any. In 2013, moreover, most confirmation laboratories did not identify 25I-NBOMe: a 25I-NBOMe patient then presented as a 2C-I case.

That is why the two published severe cases cannot be charged to 2C-I. In Thornton's 17-year-old boy, 2C-I and 25I-NBOMe were present in serum and in urine after two blotters: this is very probably an exposure to 25I-NBOMe alone. In Bosak's 19-year-old man, the abstract writes only that a comprehensive screen revealed the presence of 2C-I, which, in 2013 and given the metabolite, does not establish what he had taken.

The confusion that actually killed, and it runs the other way

Blaine, Minnesota, March 2011. A product bought on the internet and advertised as 2C-I was identified on analysis as 2C-E, the Minnesota Bureau of Criminal Apprehension finding 8.9 grams in the seized vial. Ten individuals aged 16 to 23, tachycardia, two patients with a systolic above 220 mmHg, eight in severe agitation with delirium and hallucinations. Outcomes: one death, one severe case, eight minor cases.

The arithmetic of the substitution deserves to be set out, because it explains why a labelling error is enough. The dose tables in circulation, which are user corpora and not measurements, give common 2C-I at 10 to 25 mg by the oral route and 2C-E at 10 to 15 mg, the very strong band for 2C-E starting at 25 mg. And the transposition was double, molecule and route: the product was insufflated, and on the nasal table for 2C-E the very strong band starts at 10 mg. A common dose of 2C-I taken by the nose therefore represents one to two and a half times the threshold of the very strong nasal band for 2C-E. Two visually identical white powders, and the route aggravated the molecule.

Status of these facts, not to be upgraded: the clinical figures of that evening come from a table in a peer-reviewed article, but the death itself rests, in the indexed literature, only on an unread conference abstract and on a second table in the same article, explicitly entitled deaths reported by the media and carrying the note that their accuracy cannot be verified. The sources do not count the same way either, one giving ten hospitalised in addition to the dead man, the other ten in all. The same market delivered, the same year, aniline sold as 2C-E to two Oregon men, cyanotic at fifteen minutes and treated for refractory methaemoglobinaemia: a white powder bought online is only what the seller says it is.

What the molecule does, when it really is the molecule

A 5-HT2A agonist, and dependence on that receptor is established and replicated: in the mouse as in the rat, the selective antagonist M100907 abolishes the behaviours induced by 2C-I. Beyond that, the laboratories do not agree, and it is better to show the spread than to average it. 5-HT2A affinity: 9.3 nM on recombinant human receptor with an agonist radioligand, about 125 nM on rat brain membranes with an antagonist radioligand, a derived figure but cross-checked by three concordant routes. This factor of thirteen comes down to protocol, an agonist radioligand preferentially labelling the high-affinity state. In functional potency, by contrast, it really is a disagreement: EC50 of 3.97 nM in a calcium assay, 60 nM in a comparable assay from another team.

The most singular result is a dissociation downstream of the same receptor: 2C-I is a full and very potent agonist of the phospholipase C pathway, and it antagonises the serotonin-stimulated release of arachidonic acid, alone of six phenethylamines tested. It is a single laboratory result, unreplicated, contradicted by an earlier report on CHO cells; it is an open puzzle, and it must be read as such. The same disagreement at 5-HT2C, full agonist for one team, partial at about 30 per cent efficacy for the other. 5-HT1A, at 970 nM, is probably irrelevant at human doses. 2C-I is the only one of the six compounds tested to have a measurable affinity for the serotonin transporter, in the high nanomolar range, but with a very low reuptake inhibition potency, no releasing activity at the three transporters, and no measurable affinity for DAT or NET.

It is not a stimulant. In the rat, 2C-I fully substitutes for DMT and for LSD, produces only 38 per cent appropriate responding in rats trained to methamphetamine, and in the mouse it depresses locomotor activity from 3 to 30 mg/kg, with no stimulant phase at any dose tested. Nor is it the empathogen its reputation claims: no publication on its sexual or empathogenic effects exists, and that reputation is probably inherited from that of 2C-B, which is another molecule.

Finally, the point that governs interactions. The deamination of 2C-I goes essentially through MAO-A and MAO-B. CYP2D6 does not appear for this compound, but nor does it appear for 2C-B, and where it does intervene the authors write that it intervenes only to a very small extent: there is no backup route to lose, there is a single route that has always been single. In the other direction, the 2C compounds themselves weakly inhibit the MAOs, IC50 of 10 to 125 µM at MAO-A for the series; the line specific to 2C-I has not been read, and the authors conclude that a pharmacological contribution of this inhibition cannot be excluded. The two mechanisms can coexist.

Duration, dose, tolerance

The figures in this section are the ones that hurt, and their provenance counts as much as their value. By the oral route, the range in circulation is 14 to 22 mg for 6 to 10 hours. It appears to be cross-checked by two sources, it has only one: PiHKAL, entry 33, a 1991 self-experiment, and the toxicological review cited second almost certainly copied it, from an earlier record that nobody has been able to re-verify. Neither the dossier nor its refutation was able to read PiHKAL first-hand, nor to reopen the table in that review. The user tables, for their part, give a wider common band, 10 to 25 mg. The often quoted duration of 6 to 8 hours is wrong against the source. A single time marker, third-hand: onset within 40 minutes, peak around 2 hours.

In other words: a range drawn from a single self-experiment, with no human pharmacokinetics, no dose-response curve, no measure of interindividual variability. It says nothing about the margin between active dose and problematic dose.

By the nasal route, no published range exists. It is a hole, it is dangerous, and it must be written rather than filled. Do not transpose from a neighbour: the oral versus nasal kinetic figures the literature gives are for 2C-T-7, onset in 5 to 15 minutes by the nose against 1 to 2.5 hours by the mouth, and carrying them over to 2C-I would be exactly the slippage this page seeks to prevent.

On blotter, no marker either, and the medium is a signal in itself. Unit doses of 25I-NBOH on blotter run from 300 to 1000 µg, that is fifteen to seventy times less than an oral dose of 2C-I. A blotter is not a medium for 2C-I: a blotter active at a tenth of a milligram is an NBOMe or an NBOH.

On tolerance, no data, neither acute, nor after repeated administration, nor crossed with LSD or psilocybin. Borrowing the LSD curve would be an extrapolation without a source.

Harm reduction

Identity first, because that is the dominant risk. A colorimetric street reagent does not separate one 2C from another 2C, and does not separate 2C-I from 25I-NBOMe either, both phenethylamines. The only nameable method that fully differentiates 25I-NBOH, 2C-I and 25I-NBOMe is differential pulse voltammetry on a screen-printed electrode: that is the name to give to an analysis service; failing that, ask for a derivatisation, or a method that does not expose the sample to a hot injector. On a blotter, bitterness and numbness of the mouth signal an NBOMe, and that is the last moment at which the exposure can be interrupted.

MAOIs are the best-founded interaction on this page, and it has no case. 2C-I is eliminated by the MAOs; blocking them closes its exit route and makes the dose unpredictable. The dominant mechanism is an overdose by blocked clearance, not first of all a serotonin syndrome. This concerns the prescribed MAOIs (moclobemide, phenelzine, tranylcypromine, selegiline, rasagiline), those not recognised as such (linezolid, methylene blue), and in this batch Syrian rue, a reversible and competitive inhibitor of MAO-A, as well as ayahuasca and Banisteriopsis caapi; the argument that a reversible inhibitor would keep one safe does not hold. AMT is the second established MAO inhibitor of the batch, and it is a serotonin releaser besides: the same reasoning applies to it with less force, its inhibitory potency not being quantified, the only range read covering three orders of magnitude. A query designed so that the answer could be seen, crossing the synthetic psychedelics of this batch with the MAOIs and harmala, returns seven results and no interaction case. The absence is real, it is not a query defect. It makes this combination a thing to avoid by reasoning, and it leaves the clinical file empty: the two halves of that sentence must stay together.

On serotonergic antidepressants, no published case associates 2C-I with an SSRI or an SNRI. The mechanism is plausible, 5-HT2A agonism plus a weak but measurable SERT affinity, plus a MAO-dependent clearance. This remains a plausibility, and it presents itself with far less force than the MAOI interaction.

With MDMA, cocaine and the stimulants, there is neither case nor study for 2C-I, and a danger must not be manufactured to fill the box. What exists is a single polydrug case, a 39-year-old woman brought to the emergency department after cocaine, reported MDMA and 2C-I, urine analysis finding 2C-I and MDA and ruling out MDMA: haemorrhagic stroke and quadriplegia, on a Moyamoya-type cerebrovascular anomaly that the authors describe as secondary to substance use. Cocaine is present, no causal link with 2C-I is established, and the case does not exonerate anything either. The only precedent of scale in this batch concerns another molecule: the DEA attributes two of the three 2C-T-7 deaths to concomitant MDMA use.

Seizures are what is best documented about 2C-I in humans, and it is there that it joins almost all of this batch. The two severe cases published under its label are seizure states, one recurrent, the other followed by apnoea and intubation: 2C-I is therefore one of the six whose seizure signal rests on counted human cases, and not on an animal signal, an isolated case or an unquantified clinical mention. Ten of the fourteen substances of this batch carry this signal, and in its case it is among the best supported: mixing it with one of the nine others, or with tramadol and lithium from outside the batch, lowers the same threshold twice, and lithium is the only medicine to which a death in this batch has been imputed.

Rhabdomyolysis can appear after the end of the psychic effects. Thornton's patient had a creatine kinase at 14,579 U/L seventy-two hours after taking it, normal on admission, and a creatinine risen to 2.46 mg/dL. To be watched beyond the night: dark urine, diffuse muscle pain, a fall in urine output, a fever that persists. Going home with no symptom does not close the episode. This case is very probably an exposure to 25I-NBOMe, which takes nothing away from the warning, since anyone who has taken what was sold to them as 2C-I carries the same identity uncertainty.

Signs that must prompt a call for emergency help without delay: a seizure even if brief, hyperthermia, rigidity, clonus or hyperreflexia, marked hypertension and tachycardia, and above all extreme agitation progressing to agitated delirium, an element that recurs constantly in the deaths attributed to the 2Cs.

In the emergency department, no one will be identified. 2C-I is not seen by a urine amphetamine screen, and the proof is direct: Thornton's patient was seizing, had 2C-I and 25I-NBOMe in serum and in urine, and his urine immunoassay screen came back negative. This wall is common to the whole batch, in both directions: 2C-T-7 has a cross-reactivity below 0.4 per cent on nine immunoassays, tryptamines such as AMT, DPT and 5-MeO-DiPT appear in no routine analysis, and diphenhydramine, conversely, produces false positives including for phencyclidine. Saying what was taken and keeping the bag materially changes the management. That management, as the clinical review describes it, is symptomatic and supportive, with no antidote: for agitated delirium, rapid sedation, fluid resuscitation and correction of the hyperthermia, benzodiazepines, ketamine cited as a rapid-onset alternative. And one formula to be kept as it stands, because the reflex is exactly the wrong one: antipyretics have no place in the treatment of a hyperthermia of toxic origin.

The risk is increased, by inference and not by data of its own, in anyone on a prescribed or botanical MAOI, in the face of any history of seizures, any known cerebrovascular anomaly, any cardiac, renal or muscular pathology, and in the case of close-spaced repeated use on 5-HT2B grounds. On this last point one must stay exact: 5-HT2B agonism is indeed the established mechanism of drug-induced valvulopathies, but the activation potencies measured in this family are 11 to 146 times weaker than at 5-HT2A and the efficacy only partial, the value specific to 2C-I has not been read, and no case of valvulopathy associated with a 2C compound exists. That justifies spacing doses out and stops there, the severity not being demonstrated.

Finally, do not dose by eye: there exist no lethality data in the rodent for 2C-I, which has never been tested above 30 mg/kg when two of its neighbours have been up to 100. Absence of testing, not a reassuring result.

What does not exist

This list is a result, not a confession. No human pharmacological study of 2C-I, controlled or observational, whereas its two immediate neighbours, 2C-B and 2C-E, have one each with a salivary assay. Nothing either in pharmacokinetics, in imaging, in EEG or in tolerance. And not one report of a death attributed to 2C-I, a query redone twice on independent databases.

This last absence is the one that must above all not be read as harmlessness. The molecule is not widespread, rarely looked for, invisible to routine screens, analytically confounded with its derivative, and above all: a 25I-NBOMe death in which 2C-I is detected will be attributed to the parent compound. The absence is therefore expected, and it means even less than it appears. The only human datum that exists is self-reported: 5.1 per cent lifetime use reported among 682 young adults attending electronic music events in New York, which makes 2C-I the most reported 2C compound. That figure measures what people believe they have taken.

Sources

• A.J. Eshleman et al. (2013), behavioural and neurochemical pharmacology of six substituted phenethylamines, Psychopharmacology (central source: binding, function, discrimination, locomotion; tables not rendered by the text retriever).
• J.S. Elmore et al. (2018), comparative neuropharmacology of NBOMe hallucinogens and their 2C counterparts in the rat, Neuropharmacology (EC50 of 3.97 nM, ED50 in the rat, potency ratios).
• A.J. Eshleman et al. (2018), psychoactive N-benzylphenethylamines, Biochemical Pharmacology (triplet 970 / 9.3 / 10.2 nM; 5-HT2B and valvulopathy framing, measured on the NBOMes and NBOHs, not on 2C-I).
• A.L. Halberstadt and M.A. Geyer (2013), effects of 2C-I and its N-benzyl derivatives on the head twitch response, Neuropharmacology (abstract only).
• D.S. Theobald and H.H. Maurer (2007), isoenzymes involved in the deamination of designer drugs of the 2C series, Biochemical Pharmacology (abstract only, re-verified word for word): primary source of the MAOI interaction.
• L. Wagmann et al. (2019), interactions of 2C series substances with human monoamine oxidases, Drug Testing and Analysis (abstract only, the line specific to 2C-I has not been read; conclusion: a contribution of MAO inhibition cannot be excluded).
• M. Hermanns-Clausen et al. (2017), bad trip on 25I-NBOMe, Clinical Toxicology (abstract only): establishes that 2C-I is an in vivo human metabolite of 25I-NBOMe and that 25I-NBOMe is active by the oral route.
• A. Bosak, F. LoVecchio, M. Levine (2013), recurrent seizures and serotonin syndrome after ingestion of 2C-I, J Med Toxicol (abstract only, empty PMC body; the dose remains unknown).
• S.L. Thornton, S. Hoehn, R.R. Gerona (2018), seizures, systemic inflammatory response and rhabdomyolysis, Pediatric Emergency Care (abstract only): blotter, CK at 14,579 U/L at 72 h, negative urine screen, 2C-I and 25I-NBOMe both present.
• B.V. Dean, S.J. Stellpflug, A.M. Burnett, K.M. Engebretsen (2013), 2C or not 2C, a review of synthetic phenethylamines, J Med Toxicol: source of the dose tables, of the distribution of deaths, of the Blaine affair and of the clinical management. Degraded status: read by the researcher, not re-verified, PMC returning a reCAPTCHA page. Its table of deaths is explicitly built on press sources, and the abstract contains no dose figure.
• J.C. Drees, J.A. Stone, A.H.B. Wu (2009), morbidity involving MDA and 2C-I, J Forensic Sci (abstract only): the Moyamoya case, on cocaine and MDA polydrug use; the authors describe the Moyamoya as secondary to substance use, this case establishes nothing and excludes nothing for 2C-I.
• A.F. Belchior de Andrade et al. (2020), identification of thermolabile substances, the case of 25I-NBOH, Forensic Sci Int (abstract only): injector degradation read as 2C-I.
• A.F. Belchior de Andrade and J. Gonzalez-Rodriguez (2019), electroanalytical identification by differential pulse voltammetry, The Analyst (abstract only): the nameable analytical countermeasure.
• L.C. Arantes et al. (2017), 25I-NBOH identified in blotter seizures in Brazil, Forensic Toxicology: unit doses on blotter of 300 to 1000 µg.
• F. Schifano et al. (2021), novel psychoactive substances and serotonin syndrome, a systematic review, Exp Neurol (abstract only).
• P. Nikolaou et al. (2014), 2C-I-NBOMe, an N-bomb that kills with Smiles, Drug Chem Toxicol (abstract only), cited above all as an illustration of the nomenclature confusion, its title using both names.
• E. Gil-Martins et al. (2025), mitochondrial dysfunction induced by 2C-I and 25I-NBOMe, Chem Biol Interact and E. Gil-Martins et al. (2025), sublethal concentrations in C. elegans, IJMS; V. Cocchi et al. (2020), psychoactive phenethylamines and genetic material, IJMS. Micromolar cellular toxicity, EC50 of about 120 µM and human genotoxicity from 35 µM. Three orders of magnitude separate these values from the plausible serum concentrations after 20 mg by the oral route: these studies say nothing about human neurotoxicity.
• J.J. Palamar et al. (2016), self-reported use of novel psychoactive substances in the nightlife setting, Am J Drug Alcohol Abuse (abstract only): the 5.1 per cent, self-reported.
• E. Papaseit et al. (2018), acute effects of 2C-B in humans, Front Pharmacol and E. Papaseit et al. (2020), acute effects of 2C-E, Front Pharmacol: cited to show what 2C-I does not have.
• A. Shulgin and A. Shulgin, PiHKAL, entry 33, Transform Press, 1991. Text not read, Erowid returning a 403 and the isomerdesign mirror being rendered by script. Primary and probably sole source of all the 2C-I dose and duration ranges in circulation. A human reading remains necessary.
• A. Rickli et al. (2015), interaction profiles of the NBOMe derivatives of the 2C phenethylamines, Neuropharmacology 99, 546-553. Not read, paywall. The 2C-I line of its table, in particular for 5-HT2B and TAAR1, is the central missing figure of this work.
• EMCDDA risk assessment report on 2C-I, 2C-T-2 and 2C-T-7, submitted to the Council on 4 April 2003, and Council Decision 2003/847/JAI of 27 November 2003. Not read: a PDF for the first, a page returning an empty body for the second. French legal status is not established here first-hand.
• Erowid, 2C-I vault. Not consulted, HTTP 403 on automated access. A corpus of user reports, to be read on site: its licence forbids reproduction, and no report is copied here. The dose bands cited above come from tables of this type, which are user compilations and not measurements.