Tianeptine Free Acid Reference Standard — CAS 72797-41-2, C₂₁H₂₅ClN₂O₄S

1. What this material is

This page describes tianeptine free acid supplied as an analytical reference material: a weighed quantity of one identified chemical substance, intended to serve as the fixed point against which another sample is measured. A reference material has a narrow job. When a forensic laboratory reports that a seized powder contains tianeptine, when a toxicology laboratory reports a plasma concentration, when a quality-control laboratory reports the content of a tablet, every one of those numbers inherits the identity, the form and the purity of one vial standing upstream of the instrument.

For most substances in a catalogue the identity question is answered by a CAS number and a formula, and the interesting analytical content lies elsewhere. Tianeptine is not one of those substances. Here the identity question is itself the difficult part, and it is difficult in a specific, arithmetically quantifiable way that no purity figure can express: the same name, the same international nonproprietary name and frequently the same catalogue entry cover two chemically distinct articles whose molar masses differ by 21.982 g·mol⁻¹. Everything in sections 3 and 4 follows from that single fact, and it is the reason this card runs long.

Three further properties compound the problem, and each is measured rather than asserted in the sections below. The public spectral record for this molecule is lopsided: the vibrational spectra that exist belong to the salt, not to the acid, and no nuclear magnetic resonance spectrum exists for either form (section 9). The routine mass-spectrometric identification is not straightforward at ordinary source conditions, because at a high cone voltage the protonated molecule is not the base peak (section 6). And the legal position is genuinely heterogeneous across jurisdictions, in a pattern that cannot be inferred from any single national list, and in which the closest structural analogue of this molecule behaves in exactly the opposite way (section 13).

Those four facts together describe a substance for which laboratories carry real, documented analytical demand. There is a sustained methodological literature running from liquid chromatography with fluorescence detection [21] through differential-pulse and adsorptive stripping voltammetry [24], ion-pair spectrophotometry [26] and near-infrared tablet assay [25] to an enantiomeric separation described only in 2025 [16] — while the publicly available reference data are incomplete precisely on the side of the form sold as a reagent. That gap is what a characterised standard closes.

For the general terms on which this shop supplies reference materials, the surrounding reference standards category collects the rest of the catalogue on the same basis.

2. Identity and registry codes

Every identifier below is quoted from a public registry, with the registry named. Where the free acid and the sodium salt carry different values, both columns are given side by side, because presenting one column alone is exactly the failure mode this page exists to prevent.

Machine-readable descriptors

3. Free acid and sodium salt: two articles, one name

This is the analytical heart of the card, and it deserves stating in the bluntest available terms.

Five per cent is not a rounding matter. It is wider than the acceptance window of most content-uniformity specifications, wider than the accuracy criterion of most validated bioanalytical methods, and comparable in size to the entire permitted impurity load of a typical pharmaceutical substance. A laboratory that calibrates with the acid and reports against a salt-referenced specification, or the reverse, produces results that are internally consistent, fully documented and biased by an amount no system-suitability test will flag — because nothing in the chromatography is wrong. The chromatography is fine. What has failed is the arithmetic that converts peak area into mass.

Why the confusion is structural rather than accidental

It would be comfortable to treat this as a careless-supplier problem. The measured record does not support that reading. The ambiguity is present in the registries themselves, at four separate levels.

First, in the synonym lists. The removed-synonym list of the free-acid record CID 68870 includes the string 30123-17-2 — the CAS number of the sodium salt — together with tianeptine monosodium and Tianeptineacid [1]. Those entries were removed precisely because they were wrong, but their presence documents that the salt number has at some point been attached to the acid record by an upstream data source. A purchaser who validates a supplier certificate by matching the CAS number alone can therefore match successfully against the wrong form.

Second, in the state register itself. Of the six authorised Polish medicinal products, five carry Tianeptinum natricum in both the common-name field and the active-substance field. The sixth — authorisation 03702, held by Les Laboratoires Servier — carries Tianeptinum natricum as the common name and the bare Tianeptinum as the active-substance name [9]. This is a national medicines register, maintained under statutory obligation, recording the form inconsistently inside a single record. If that register does not hold the distinction uniformly, a commercial catalogue certainly will not.

Third, in the pharmacopoeial marker. Among the deposited synonyms of the salt record sits the string TIANEPTINE SODIUM [EP MONOGRAPH] [2]. If that marker is accurate, the European Pharmacopoeia monograph is written against the salt, which would mean a laboratory testing to a pharmacopoeial method needs salt-referenced quantities while holding acid-referenced material. We flag this as a marker present in an aggregated synonym list, not as a verified pharmacopoeial fact: it has not been confirmed against the issuing authority, and it should not be relied upon without that confirmation. The wider point survives either way, because the six Polish authorisations independently establish that the medicinal form is the salt [9].

Fourth, in the pharmacokinetic and formulation literature. The formulation papers name the salt explicitly — a sustained-release dual-layer tablet study [29], an orodispersible film [30], a bioequivalence assessment of two formulations [31] and a stability-indicating chromatographic assay [22] are all written on tianeptine sodium. The bioanalytical and pharmacokinetic papers, by contrast, generally name tianeptine without qualification [20][21][27]. Reading across those two bodies of work without tracking the form is how a five-per-cent bias enters a comparison silently.

Other salts that exist

The acid and the monosodium salt are not the only registry entries with this connectivity. Two further salt forms are recorded, and both would fail a mass-based identity check against either of the main forms:

Note the hydrochloride in particular. It contains two chlorine atoms rather than one, which changes the isotopic envelope in a way that is directly visible in a mass spectrum: a monochloro compound shows an M+2 satellite at roughly one third the height of the monoisotopic peak, a dichloro compound one at roughly two thirds. Section 6 returns to that pattern, because for this molecule the chlorine doublet is among the most useful confirmatory features available.

4. Identity traps in the numbers themselves

Beyond the salt question, the identifiers attached to this molecule carry three further traps. Each is documented in the registry record rather than inferred, and each has a concrete consequence for how a purchase order or a certificate should be written.

Trap one: two CAS numbers, one deprecated, both in circulation

The free-acid record lists two CAS Registry Numbers drawn from different reporting sources [1]:

This is not a case where the older number simply fell out of use. It is still emitted by live registry sources, it still has its own EC number attached, and it circulates widely in commercial listings. A laboratory searching an internal archive on 72797-41-2 alone will miss records filed under 66981-73-5, and a purchasing system that validates a certificate against a single expected CAS number will reject correct material. Search on both; write the current one on documentation.

The enantiomers carry their own registry numbers again, which matters for anyone specifying material for chiral work: 169293-31-6 for the (+) enantiomer and 191172-75-5 for the (−) enantiomer, with unique-ingredient identifiers AKU7QFL9ZT and XV6773012I respectively [1].

Trap two: the classification of the molecule contradicts itself between vocabularies

The medical subject heading vocabulary places tianeptine among Antidepressive Agents, Tricyclic. The anatomical therapeutic chemical code assigned to it is N06AX14 [15] — which sits in N06AX, other antidepressants, and specifically not in N06AA, the tricyclic group. Both classifications are attached to the same registry record [1].

Structurally, the second is the better description of what the molecule is: the core is a dibenzothiazepine S,S-dioxide, not a dibenzazepine or dibenzocycloheptene, and it carries a seven-carbon aminoheptanoic side chain that no tricyclic antidepressant possesses. The consequence for an analyst is practical rather than semantic: a screening panel assembled by therapeutic class from the medical subject heading vocabulary will file this compound with a group whose chromatographic behaviour and fragmentation it does not share, and a library search restricted to that group will be searching the wrong shelf.

Trap three: the closest structural analogue is a different, and differently regulated, substance

Amineptine is the compound this molecule is most likely to be confused with on paper, and the confusion runs in both the chemical and the legal direction. It carries the same 7-aminoheptanoic side chain on a tricyclic core; the difference is that its core is a dibenzo[a,d]cycloheptene rather than a dibenzothiazepine dioxide. That single substitution changes the formula from C₂₁H₂₅ClN₂O₄S to C₂₂H₂₇NO₂ and the molecular mass from 436.951 to 337.463 [5].

Read the last four rows together. The two molecules are structural cousins whose regulatory positions are mutually inverted in every jurisdiction measured, and neither has a deposited condensed-phase spectrum of any kind. Section 13 sets out how that inversion was measured, including the positive controls that make each negative statement meaningful.

5. Structure and stereochemistry

Tianeptine is built on a tricyclic dibenzo[c,f][1,2]thiazepine core bearing a sulfone, an N-methyl group on the ring nitrogen and a chlorine on one aromatic ring, with a seven-carbon aminoheptanoic acid chain attached through a secondary amine at position 11. The stereodescriptor counters are quoted verbatim from the registry record, with the counters of the two resolved enantiomers alongside for comparison [1][2][3][4]:

The reading is unambiguous. There is exactly one atom stereocentre, zero bond stereocentres, and in the record corresponding to the commercial substance the single centre is undefined. Undefined, in registry terms, means the configuration is not specified, which for a substance sold as a single article means a racemate. The two resolved enantiomers have the same counter set with the definition moved from the undefined column to the defined column, which is exactly what a resolved single isomer looks like.

The stereocentre is the carbon at position 11 of the benzothiazepine ring — the one carrying the secondary amine that begins the heptanoic chain. That is visible directly in the systematic names of the resolved forms, which are written with the (11S)- and (11R)- prefixes.

Two independent statements in the same aggregated record confirm the racemic reading rather than merely restating it. The chemical-entities description quoted in the record reads: «Tianeptine is a racemate comprising of equimolar amounts of (R)- and (S)-tianeptine» [1]. And the naming used by the European chemicals inventory carries the racemic prefix explicitly: «(±)-7-[(3-chloro-6,11-dihydro-6-methyldibenzo[c,f][1,2]thiazepin-11-yl)amino]heptanoic acid S,S-dioxide» [8]. Three sources, three vocabularies, one conclusion.

Why the sulfone matters more than it looks

The two sulfone oxygens are the single most consequential feature of this molecule for analysis, and they are easy to overlook because they sit in the middle of the ring system rather than at a reactive periphery. Three consequences follow directly, and all three are used later on this page.

First, the sulfone accounts for a large share of the topological polar surface area — 95.1 Å² for a molecule of only 29 heavy atoms [1] — which together with the free carboxylic acid explains why the computed partition coefficient is only 1.2 despite the extended aromatic system and the seven-carbon chain. Section 7 turns that number into a chromatographic prediction.

Second, the sulfone is the origin of the diagnostic 63.9619 Da neutral loss that dominates the fragmentation pattern in section 6. Very few structural motifs give that exact loss, which makes it a strong confirmatory feature rather than a generic one.

Third, the sulfone together with the carboxylic acid is why the boiling point quoted in section 8 is physically meaningless: a molecule carrying both an aryl sulfone and a carboxylic acid decomposes long before any distillation temperature is reached, so a predicted boiling point above 600 °C describes a computational model, not a substance.

6. Mass spectrometry: the molecular ion is not the base peak

Mass spectrometry is the technique most laboratories will reach for first, and for this molecule it works — but not in the way a method transcribed from a generic template expects. The deposited spectra show a compound that fragments readily in the source, and the consequence is that the identification anchor cannot be the protonated molecule alone.

The fragmentation series, assigned and checked against measurement

The ion assignments below are computed from monoisotopic masses with the electron-mass correction applied, and set against measured values from two different deposited records: the unit-resolution Waters series for the free acid [7], and one high-resolution qTof MS² spectrum deposited under the sodium-salt record [2], which supplies the only six-decimal values quoted here:

The pathway is coherent and easy to remember: the aminoheptanoic side chain is lost first to give the tricyclic core at 292, then sulfur dioxide departs as a neutral of 63.9619 Da to give 228, then chlorine departs as 34.9689 Da to give 193, and ring contraction gives the fluorenyl cation at 165. Four ions, three well-defined neutral losses, one chain.

Note what that pathway implies about specificity. The side chain leaves first. That means the ions at 292, 228, 193 and 165 are shared with anything that carries the same tricyclic core — including, by construction, the beta-oxidation metabolites in section 10, whose only difference from the parent is the length of that chain. Those fragments confirm the core; they do not distinguish the parent from its metabolites. The distinction lives in the precursor mass, which returns the burden to the molecular ion this section opened by warning about.

The chlorine doublet is the most useful free confirmation available

Chlorine has two stable isotopes in a roughly three-to-one abundance ratio, and this molecule carries exactly one chlorine atom. Every chlorine-containing ion in the series therefore appears as a pair separated by two mass units with the satellite at approximately one third the height of the main peak. The deposited spectra show the pairs plainly: 437/439 in the molecular-ion region, 292/294 for the core, 228/230 after sulfur dioxide loss [7]. The pair disappears at 193, which is precisely the point at which chlorine has been lost — internal confirmation that the assignment is right.

This gives a laboratory three independent checks that cost nothing extra: the count of chlorine atoms in the molecule, the identity of which fragments retain the halogen, and a quick discriminator against the dichloro hydrochloride salt form in section 3, whose satellite would sit at roughly two thirds rather than one third of the main peak height.

All seven MassBank records — WA000311 through WA000317 — were acquired on the same chromatographic system, a 2.1 mm, 3.5 µm C18 column with a retention time of 13.560 min, and are distributed under a non-commercial licence [7]. They are not seven independent measurements but one sampling-cone ladder run on one material: 90, 75, 60, 45, 30 and 15 V in positive mode, plus 30 V in negative mode. That single-laboratory provenance is worth noting for the same reason it is noted in section 9: seven spectra from one source are seven observations of one material, not seven independent confirmations.

7. Methods that work, methods that fail, and why

The computed descriptors in section 2 are not decoration; they predict chromatographic behaviour, and the predictions are testable against the published method literature. Reading them together gives a usable map of which approaches will work on this compound and which will disappoint.

What the published methods actually are

The methodological literature on this compound is broader than for most substances of comparable obscurity, and it clusters into five families. Naming them is useful, because it tells a method developer where the prior art is.

Liquid chromatography with fluorescence detection is the best-established route for plasma. A determination in human plasma by that technique was published in 2006 [21], following an earlier reversed-phase method with conventional detection applied to pharmacokinetic work in 2000 [20]. A stability-indicating assay for the sodium salt runs fluorescence and ultraviolet detection simultaneously [22], which is the direct answer to the ultraviolet-reference gap noted above — the two channels cross-check each other within a single injection.

Tandem mass spectrometry is the modern route, and the reference method in the literature was developed to quantify the parent together with its active metabolite in a single run across several administration routes [27]. That paper matters for reasons beyond its own results: it establishes that parent and metabolite must be resolved chromatographically, since as section 6 showed they share their diagnostic fragments.

Electroanalysis is documented for tablets by differential pulse, square wave and adsorptive stripping voltammetry [24]. This is an unusual entry in a modern method list and it is a genuinely useful one: it needs no chromophore, no column and no organic solvent, and it is the kind of approach that survives when the reference spectra are missing.

Spectrophotometry and spectrofluorimetry cover the simple-laboratory end. An ion-pair spectrophotometric method for tablets was published in 2008 [26], and a spectrofluorimetric determination exploiting a quenching interaction appeared in 2021 [23].

Near-infrared spectroscopy has been applied to the content of coated tablets, with explicit treatment of model optimisation, calibration transfer between instruments and confidence intervals [25]. Calibration transfer is the operationally interesting part: it is the problem of making one instrument agree with another, which is the same problem a reference material solves in the chemical domain.

8. Physicochemical data, attributed value by value

The experimental-properties section of the free-acid record contains three entries in total [1]. That is the whole of the measured physical record for this substance in the aggregated registry, and each of the three carries a caveat.

What is absent, and why the absence is an established finding

The following quantities have no experimental value in the record for the free acid [1]: aqueous solubility, solubility in any organic solvent, acid dissociation constant, measured partition coefficient, density, decomposition temperature, hygroscopicity, and any description of polymorphic forms. The only partition coefficient available anywhere in the record is the computed XLogP3-AA value of 1.2, which is an estimate and is labelled as one.

For the sodium salt the position is starker still: the record for CID 23663953 has no experimental-properties section at all [2]. The form that is the active substance of every medicinal product authorised in Poland under this name [9] has no measured physical quantity of any kind deposited in the aggregated registry — not a melting point, not a solubility, nothing.

This is an unusual shape for a data record, and it is worth being precise about what it means. It does not mean that nobody has measured these things; formulation work of the kind published on the salt [29][30] cannot be done without solubility data. It means those measurements live in the primary literature and in unpublished development files, and have never been abstracted into the public registries. The practical consequence for a laboratory is identical either way: there is no authoritative public value to compare your own measurement against.

9. Spectra: what exists, what does not

Spectral coverage is where the public record for this molecule is at its thinnest, and the shape of the gap is the single most operationally important fact on this page after the salt question. The measurement below was made by querying the spectral-information section of the registry for each record, with three substances of known-rich coverage included in the same query as controls.

The controls fire. The same query, run in the same way against the same registry, returns nuclear magnetic resonance and ultraviolet headings for all three control substances. The zeros in the first three rows are therefore a finding about those substances, not an artefact of an instrument pointed at nothing.

Three consequences, each uncomfortable in a different way

First: there is no nuclear magnetic resonance spectrum for either form. Not proton, not carbon-13, not two-dimensional. For a molecule whose identity question is a question about a proton — whether the carboxyl group carries a hydrogen or a sodium — this is the missing measurement that would settle the matter most directly. A proton spectrum distinguishes the acid from the salt immediately, through the presence or absence of the exchangeable carboxylic proton and through the chemical shift of the methylene group adjacent to the carbonyl. Nobody has deposited one.

Second: the vibrational spectra that exist belong to the salt, and to one lot of it. Infrared in potassium bromide, attenuated-total-reflectance infrared, and Fourier-transform Raman are all deposited for the sodium salt [2]. All three were recorded on a single sample from a single supplier lot — catalogue item 17561, lot 0467324-1 — deposited by a forensic spectral research group. Three techniques applied to one material tell you three things about that material; they do not tell you that a second lot would look the same. And since the deposit is of the salt, a laboratory confirming the identity of the acid by infrared has no public reference spectrum at all. The carboxylic acid and the carboxylate differ sharply in exactly the region an analyst would look at: a free acid shows a carbonyl stretch near 1700 cm⁻¹ and a broad hydroxyl envelope, a carboxylate shows the paired asymmetric and symmetric stretches near 1560 and 1400 cm⁻¹ and no such envelope. Comparing an acid sample against the deposited salt spectrum will show a mismatch that is entirely correct and entirely misleading.

Third: there is no ultraviolet reference spectrum, while a substantial part of the quantitative literature uses ultraviolet detection [22][26]. The molecule has a perfectly serviceable chromophore; this is a gap in deposition, not in physics. It nonetheless means that the absorption maximum and the absorptivity used to build a calibration have to be taken from a method paper or measured in-house, and that two laboratories quoting different wavelengths for this compound cannot resolve the difference by appeal to a registry.

One more point about that provenance. The vibrational spectra come from a forensic depositor, not from a pharmacopoeial one. That tells you something about who has been characterising this compound and why: the reference material was made and measured to support the identification of seized samples, which is consistent with the substance appearing in the national forensic laboratory reporting system since 2018 (section 13). It is a useful spectrum and a legitimate source. It is not a pharmacopoeial reference, and it should not be cited as one.

10. Metabolites and degradation products

Two families of related compounds matter to anyone using this material: the metabolites, because they appear alongside the parent in any biological sample and share its diagnostic fragments; and the photodegradation products, because they bear on how the standard itself must be stored.

Beta-oxidation shortens the side chain in two-carbon steps

The registry record states the metabolic route plainly, quoting an aggregated pharmacology source: metabolism proceeds «primarily by beta-oxidation of its heptanoic side chain», and «three major metabolites result» [1]. Beta-oxidation removes C₂H₄ units, so the products are a homologous series differing by exactly 28.0313 Da each step, with the tricyclic core untouched.

MC5 is described in the title of the reference bioanalytical paper as an active metabolite [27], and that paper exists precisely because parent and metabolite have to be quantified together. Section 6 explains why they cannot be told apart by fragmentation alone: the side chain is the first thing lost in the source, so parent and both metabolites converge on the same core ion at m/z 292 and the same downstream series at 228, 193 and 165. The only mass-domain discriminator is the precursor. Chromatographic resolution is therefore not a nicety in this analysis; it is the entire basis of the separation between the three.

Pharmacokinetic parameters, as description of the substance

The aggregated record carries a set of pharmacokinetic figures, reproduced here as bibliographic description of the molecule and not as information about this article or any use of it [1]: oral bioavailability approximately 99 per cent; plasma protein binding approximately 95 per cent; volume of distribution 0.8 L·kg⁻¹ (reported as 0.77 ± 0.31); elimination half-life approximately 2.5 hours; biliary elimination as glucuronide and glutamine conjugates; approximately 66 per cent of a radiolabelled quantity recovered renally over one week, with approximately 3 per cent unchanged in urine. The short half-life and the extensive conjugation are the reason the metabolite question in this section is an analytical problem rather than a footnote.

Photodegradation, and what it means for storage

A non-targeted study published in 2024 identified the photodegradation products of tianeptine in water by ultra-high-performance chromatography coupled to quadrupole time-of-flight tandem mass spectrometry [28]. The work was environmental in motivation, but its findings are directly relevant to a reference material: they establish that this molecule is photolabile in aqueous media and that the transformation gives multiple products rather than one.

Read that alongside the stability-indicating chromatographic assay developed for the sodium salt [22], which exists because the compound needs a method capable of separating it from its own degradation products. The two papers together justify the storage guidance in section 14 without any need to extrapolate: protect from light, store dry, and do not leave aqueous stock solutions standing. No formal stability study on this material as a reference standard exists publicly, and we will not invent one; the photodegradation finding is the reason the precaution is stated as a documented requirement rather than a habit.

11. What the literature actually says

This section reports what has been published about this molecule. Every statement in it is a bibliographic fact about the literature — that a paper exists, in a named journal, reporting a named finding. None of it describes this article, and none of it is a claim about any use of this reagent.

An unusual pharmacology for the class, extensively documented

The molecule carries an anatomical therapeutic chemical code in the antidepressant group [15], and the older clinical and neurobiological reviews describe it in those terms [38][39]. Beginning in 2014, a separate body of work reported that the compound acts as an agonist at the mu-opioid receptor [32]; subsequent studies reported that its behavioural effects in animal models require that receptor [33], localised the relevant receptors to hippocampal inhibitory interneurons [34], described opioid-like adverse effects in rodents [36] and reported opioid-receptor-dependent oscillatory activity in the rat hippocampus [37]. A human-model study examined respiratory effects in the framework used for opioid-induced respiratory depression [35]. A narrative review published in 2023 collected the pharmacology and the abuse-potential literature together [55], and a 2024 review article in the Classics in Chemical Neuroscience series covered the compound's history and chemistry [19].

The reason this matters on a reference-material card is narrow and specific: a pharmacology that does not match the classification is what drives forensic and toxicological demand for an authentic standard. Laboratories that would never otherwise stock an antidepressant standard have reason to hold this one.

The forensic and toxicological record

There is a substantial published record of poisoning-centre reports, emergency presentations and fatal cases involving this substance, spanning two decades and several countries. A fatal intoxication was reported in 2007 [40]; a series of fatalities in the United States followed in 2018 [41]. In the same year, an analysis of exposures reported to a national poison data system between 2000 and 2017 appeared in a public-health weekly report [42], alongside a poison-centre experience paper [43]. Later series extended the record: exposures reported to United States poison centres between 2015 and 2023 [44], and emergency-department visits, fatal overdoses and substance seizures in one state between 2021 and 2023 [45]. Acute toxicity following intravenous use has been reported [46], as have dependence and withdrawal in case reports and reviews [47], including neonatal presentations following exposure in pregnancy [48][49].

Two systematic reviews have appeared very recently — one on misuse, withdrawal, toxicity and clinical management published in 2026 [50], and one in 2025 that analysed the phenomenon and its public perception [51]. A separate study analysed descriptions of use appearing on social platforms [52], and a French study used a reimbursement database with a doctor-shopping indicator as a quantitative measure of misuse [54].

Where the substance is monitored without being scheduled

Independently of any control schedule, the substance is carried as a monitored entry in the United States national forensic laboratory information system, recorded in the class antidepressants with an addition date of June 2018 [1][12]. That system reports what forensic laboratories identify; it is a record of analytical activity, not a control schedule, and the distinction is set out carefully in section 13. For present purposes it is the most direct available measure of the thing this card is about: laboratories are actively identifying this compound, and identification requires a standard.

The registry record additionally carries a dietary supplements section referencing a national label database, and a classification indicating the substance appears among ingredients declared on such labels [1]. It is also flagged as positive in a liver-toxicity knowledge base, and its development status is recorded as phase 3 in a curated bioactivity database [1]. All four are recorded here as registry facts about the molecule.

12. Hazard classification, and how thin it is

The classification data below are quoted as aggregated in the compound record from notifications made to the European classification and labelling inventory [1][2][8]. The number of notifications behind each classification matters at least as much as the classification itself, and for this substance those numbers are small.

How should a laboratory act on a classification that disagrees with itself? The defensible answer is the conservative one: plan the risk assessment against the most severe statement notified for either form — that is acute toxicity category 3 by the oral route, together with the reproductive, lactation and repeated-exposure statements — and record explicitly in the assessment that the classification is unharmonised, that it rests on one notification for the acid and four for the salt, and that one of those four disputes classification entirely. A percentage figure in this inventory is not a measurement of agreement among experts; it is a count of filings, and with a denominator of one or four the arithmetic is fragile.

One further note on provenance. The classification and registration data above were read from the aggregated compound record rather than from the inventory interface directly, and are attributed accordingly [1][2][8]. That distinction is worth preserving on a card of this kind: the values are reported as they are aggregated, and anyone relying on them for a regulatory filing should confirm them against the inventory itself rather than against this page.

13. Regulatory status, measured jurisdiction by jurisdiction

This is the section where a general statement would be actively harmful, because for this substance no general statement is true. The position differs between countries, differs between the federal and state level inside one country, and inverts entirely when compared against the closest structural analogue. Every statement below was measured against a named document, and every negative statement is accompanied by a positive control — the same search, on the same document, for substances that are certainly present. A zero without a firing control is not evidence of absence; it is evidence of nothing.

Poland: not scheduled, while the closest analogue is

The paired control is the strongest form this evidence can take. Amineptine shares the seven-carbon aminoheptanoic side chain and the tricyclic architecture; it is the nearest structural neighbour this molecule has in the pharmacopoeia of controlled substances. The instrument that failed to find tianeptine found its closest analogue in the same document, in a named group, at a named item number. The zero is therefore a measurement rather than a silence.

A second control establishes that the schedule is actively maintained rather than dormant. The July 2026 amendment added carisoprodol to psychotropic group IV-P at item 81; hexahydrocannabinol, 2-methylmethcathinone and N-ethylnorpentedrone to group II-P at items 83 to 85; and five nitazene-type opioids to narcotic group I-N at items 212 to 216, transposing a delegated directive of the European Union adopted in 2025 [11]. Two further entries, items 67 and 68, were added to the annex of new psychoactive substances, and two existing entries were repealed. A list that gained eleven entries and lost two in its most recent revision is a live document, and this substance is not among the entries it gained.

Poland: authorised as a medicine, in the salt form only

The second Polish measurement runs in the opposite direction and returns a positive. Querying the national register of medicinal products on the common name returns six authorisations, all of them for the sodium salt at a strength of 12.5 mg [9]:

Three things in that table deserve explicit statement. Six of six are the sodium salt — the medicinal form in Poland is not the acid, which is the point section 3 rests on. Two of the six are parallel imports whose stated validity ran out during 2026, on 16 March and 17 August, so the count is not a permanent fact and carries a time stamp rather than a guarantee. And one record spells the form inconsistently across its own two fields, which is the inconsistency discussed in section 3.

The instrument was checked in both directions. Positive controls on the same interface and the same common-name field: sertraline returns 32 records, mirtazapine 43, buprenorphine 20, paracetamol 291. A query on the product-name field for paracetamol returns 43; the two fields are different instruments and are not interchangeable [9]. A negative control was also run and failed, which is reported here rather than concealed: a query against the active-substance field returned 22,871 records, that is to say the entire register including sodium chloride and gentian violet. That field does not function as a filter, the query was discarded as an invalid instrument, and its output was not used. Only the name and common-name queries are relied upon.

European Union

At Union level the substance is the active ingredient of nationally authorised medicinal products, several of them approved through the decentralised procedure, and carries the anatomical therapeutic chemical code N06AX14 [15][9]. It is a registered substance under the chemicals regulation, dossier 25048, status active, with the record last updated in July 2020 [8]. It does not appear among the substances added by the most recent transposition of the 2025 delegated directive, whose additions are enumerated above [11].

United States: unscheduled federally, Schedule I in at least two states

The state-level position is different, and it is where a reading based on the federal list alone goes wrong.

The placement inside each statute is worth reading, because the two states classify the same substance by reference to different neighbours: one files it among opioids, the other among depressants. That is not an inconsistency to be resolved on this page; it is a fact about how two legislatures chose to categorise the same compound, and it tells an analyst that the substance sits at a genuine taxonomic boundary.

And the inversion holds. Amineptine returns zero occurrences in both state documents [13][14], while being federally Schedule I [12]. Tianeptine is the exact reverse. Four documents, two substances, two levels of government, and the pattern is crossed at every point.

Anti-doping status: not established

14. Handling and storage

15. What we certify and what we do not

16. Terms of supply

This material is supplied as a laboratory reagent and analytical reference material, for in-vitro laboratory use by qualified personnel in an appropriately equipped facility. It is not a medicinal product, not a dietary supplement, not a food, feed or cosmetic ingredient, and it is not supplied for administration to humans or animals in any form or by any route.

By ordering, the purchaser confirms that they are a professional user acquiring the material for laboratory purposes; that they will handle it in accordance with section 14 and their own institutional risk assessment; that they will not administer it to humans or animals, nor supply it to any person who intends to; and that they are responsible for compliance with all laws applicable at the destination.

That last point carries unusual weight for this substance and should not be treated as boilerplate. Section 13 documents a position that is unscheduled in Poland, unscheduled federally in the United States, and Schedule I in at least two United States states, with the position in several further states unknown and deliberately not extrapolated. The purchaser, not the supplier, is in a position to know which law applies to a given delivery address, and the responsibility for that determination sits with them.

Nothing on this page is medical advice, nor an offer of a medicinal product. Statements in sections 11 and 13 describe research literature and medicines authorised elsewhere; they describe neither this article nor any use of it.

17. Questions and answers

Is this the free acid or the sodium salt?

The free acid: C₂₁H₂₅ClN₂O₄S, CAS 72797-41-2, relative molecular mass 437.0, monoisotopic mass 436.1223562 Da, InChIKey JICJBGPOMZQUBB-UHFFFAOYSA-N. The sodium salt is a different article with a different CAS number, a different pair of EC numbers and a mass of 458.9. If you need the salt, this is not it.

Why does the difference between the two forms matter so much?

Because it is 5.03 per cent by mass, and because nothing in the chromatography reveals it. Salt to acid is a factor of 0.95210, acid to salt 1.05031. One thousand milligrams of the acid is 1050.31 mg of salt equivalent. That bias is larger than most content-uniformity windows and larger than most bioanalytical accuracy criteria, and it survives every system-suitability test because nothing in the separation is wrong.

How do I check which form I am holding, without weighing anything?

Three ways, in increasing order of instrument requirement. Read the covalently-bonded-unit count in your data record: 1 is the acid, 2 is the salt. Read the last block of the InChIKey: -N is the acid, -M is the salt. Or run a positive-mode mass spectrum under gentle source conditions and look at the molecular-ion region: the acid gives [M+H]⁺ at m/z 437.1296, the salt gives 459.1116, and the two are 22 units apart.

Which CAS number should appear on my documentation?

72797-41-2, with EC 276-851-9. A second number, 66981-73-5 with EC 614-004-0, is carried in the same record as deprecated but is still emitted by live registry sources and circulates widely in commerce. Search your archives on both; write the current one on paperwork. And be aware that the salt number, 30123-17-2, has historically been attached to the acid record in error — it appears among the removed synonyms — so CAS matching alone does not confirm form.

Is this material chiral?

It has one stereocentre and it is racemic. The registry counters read: atom stereocentre count 1, defined 0, undefined 1, bond stereocentre count 0. The chemical-entities description in the same record calls it a racemate comprising equimolar amounts of the two enantiomers, and the European inventory names it with the plus-or-minus prefix. The centre is carbon 11, the one bearing the amine.

Can I use it to determine enantiomeric purity?

No. A racemate has no enantiomeric excess to certify and cannot establish which peak in a chiral separation belongs to which configuration. You need resolved material for that, and the enantiomers have their own registry numbers — 169293-31-6 and 191172-75-5. A full enantioselective separation with structural assignment was published only in 2025, so this is an area near the front of the literature rather than a settled routine.

Why is there no reference NMR spectrum?

Because none has been deposited, for either form. That is a measured finding rather than a search failure: the same query against caffeine, acetylsalicylic acid and paracetamol returns proton, carbon-13 and in two cases two-dimensional spectra. It is a genuine gap in the public record, and it is the gap that would most directly settle the acid-versus-salt question if it were filled.

There are infrared spectra listed. Can I use them?

With care, and probably not for the purpose you want. The infrared, attenuated-total-reflectance and Raman spectra deposited for this molecule are all of the sodium salt, and all three come from one sample of one supplier lot. Comparing acid material against a salt spectrum will show real differences in the carbonyl region — roughly 1700 cm⁻¹ for the free acid against the paired carboxylate bands near 1560 and 1400 cm⁻¹ — which are correct chemistry and a misleading identity result.

My mass spectrometer shows a base peak at 228 and almost nothing at 437. Is the material wrong?

Probably not; check your cone voltage. A deposited spectrum acquired at 90 V has its base peak at m/z 228 with 193 and 165 following, and the protonated molecule absent from the top five. A companion spectrum of the same compound under gentler conditions has 437 as base peak. Anchor identification on the whole series — 437, 292, 228, 193, 165 — together with the chlorine doublet, rather than on the molecular ion alone.

What is the chlorine doublet and why does it help?

The molecule carries exactly one chlorine, whose two stable isotopes occur in roughly a three-to-one ratio. Every chlorine-bearing ion therefore appears as a pair two mass units apart with the satellite at about one third the height of the main peak: 437 and 439, 292 and 294, 228 and 230. The pair vanishes at 193, which is where chlorine is lost. It costs nothing, it confirms the halogen count, and it distinguishes the compound from the dichloro hydrochloride salt form, whose satellite would be nearer two thirds.

Will my fragments distinguish the parent from its metabolites?

No, and this is worth planning for. Beta-oxidation shortens the heptanoic chain in C₂H₄ steps to give MC5 at 408.091056 Da monoisotopic and MC3 at 380.059756 Da. Since the side chain is the first thing lost in the source, parent and both metabolites converge on the same core ion at 292 and the same downstream series. The precursor mass is the only mass-domain discriminator, so chromatographic resolution carries the whole separation.

What is the melting point?

One source gives 148 °C, with no method stated, no range stated and no second source. Treat it as an orientation figure rather than a specification. The boiling point of 609.2 °C that appears in the same record is a computational prediction and is physically meaningless for a molecule carrying both an aryl sulfone and a carboxylic acid, which decomposes long before any such temperature.

Is it a controlled substance?

It depends entirely on where you are, and the pattern is not intuitive. Not scheduled in Poland, though its closest structural analogue amineptine is, in psychotropic group II-P. Not federally scheduled in the United States, though amineptine is, in Schedule I with drug code 1219. But it is Schedule I in Florida and in Minnesota, where amineptine in turn is absent. Several other state positions could not be read and were not extrapolated. Check the law that applies to your delivery address.

Is it on the anti-doping prohibited list?

We do not say, in either direction. The current list could not be retrieved from the issuing organisation or from eight national anti-doping organisations, and an older edition was deliberately not substituted, because a superseded prohibited list reads as authoritative while being wrong. If you have a competition-compliance question, obtain the current list directly.

What does the hazard classification actually rest on?

For the acid, one notification, which is why every statement shows at 100 per cent — one out of one is not a consensus. For the salt, four notifications that disagree with each other, including one asserting that no classification is warranted and a split between acute toxicity category 3 and category 4 on the oral route. Neither form has a harmonised classification, so nothing here is legally binding across the Union and another supplier may lawfully classify differently.

Why does this card say «not established» in several places instead of giving an answer?

Because in those places the source document could not be read, and substituting an older edition or a secondary summary would produce a statement that looks authoritative and is not verifiable. The specific gaps are the anti-doping list, the position of the European medicines and drugs agencies, the content of United States federal warnings, the state schedules beyond Florida and Minnesota, and the pharmacopoeial monograph marker. Each is named in section 15 rather than quietly omitted.

Do other products in this catalogue have the same salt-form problem?

The distinction is common; what is unusual here is that the name does not signal it. Alpha-GPC, CDP-choline and noopept each have their own identity question, set out on their own pages, and the general lesson is the same on all of them: a single confirmatory technique, and a single identifier, are rarely enough.

What purity is guaranteed?

Section 15 sets out what is and is not certified. Lot documentation accompanies the material and states the method used. We do not print a catalogue purity figure, because a purity number without a named method is not a specification — and for this substance a purity figure is read against a declared form, and this page declares one: the free acid.

References

Registry records

1. National Center for Biotechnology Information. 2026. "PubChem Compound Summary for CID 68870, Tianeptine." PubChem. https://pubchem.ncbi.nlm.nih.gov/compound/68870.
2. National Center for Biotechnology Information. 2026. "PubChem Compound Summary for CID 23663953, Tianeptine Sodium." PubChem. https://pubchem.ncbi.nlm.nih.gov/compound/23663953.
3. National Center for Biotechnology Information. 2026. "PubChem Compound Summary for CID 11704779, (S)-Tianeptine." PubChem. https://pubchem.ncbi.nlm.nih.gov/compound/11704779.
4. National Center for Biotechnology Information. 2026. "PubChem Compound Summary for CID 11640679, (R)-Tianeptine." PubChem. https://pubchem.ncbi.nlm.nih.gov/compound/11640679.
5. National Center for Biotechnology Information. 2026. "PubChem Compound Summary for CID 34870, Amineptine." PubChem. https://pubchem.ncbi.nlm.nih.gov/compound/34870.
6. National Center for Biotechnology Information. 2026. "PubChem Compound Summary for CID 128660, Tianeptine MC(5)." PubChem. https://pubchem.ncbi.nlm.nih.gov/compound/128660.
7. MassBank Europe. 2026. "Record MSBNK-Waters-WA000311, Tianeptine; LC-ESI-Q; MS; POS; 90 V" and companion records WA000312, WA000313, WA000314, WA000315, WA000316 and WA000317. https://massbank.eu/MassBank/RecordDisplay?id=MSBNK-Waters-WA000311.
8. European Chemicals Agency. 2026. "Information on Chemicals — classification and labelling inventory and registered-substance dossier 25048, EC 276-851-9." Values on this page are quoted as aggregated in reference [1] and [2]. https://echa.europa.eu/information-on-chemicals.

Regulatory documents

9. Urząd Rejestracji Produktów Leczniczych, Wyrobów Medycznych i Produktów Biobójczych. 2026. "Rejestr Produktów Leczniczych — public search interface." Queried on the common name Tianeptinum. https://rejestry.ezdrowie.gov.pl/rpl/search/public.
10. Minister Zdrowia. 2024. "Obwieszczenie Ministra Zdrowia z dnia 17 czerwca 2024 r. w sprawie ogłoszenia jednolitego tekstu rozporządzenia w sprawie wykazu substancji psychotropowych, środków odurzających oraz nowych substancji psychoaktywnych." Dziennik Ustaw 2024, poz. 1139. https://dziennikustaw.gov.pl/DU/2024/1139.
11. Minister Zdrowia. 2026. "Rozporządzenie Ministra Zdrowia z dnia 7 lipca 2026 r. zmieniające rozporządzenie w sprawie wykazu substancji psychotropowych, środków odurzających oraz nowych substancji psychoaktywnych." Dziennik Ustaw 2026, poz. 934. https://dziennikustaw.gov.pl/DU/2026/934.
12. United States Drug Enforcement Administration, Diversion Control Division. 2026. "Controlled Substances — Alphabetical Order." https://www.deadiversion.usdoj.gov/schedules/orangebook/c_cs_alpha.pdf.
13. Florida Legislature. 2026. "Florida Statutes, Section 893.03 — Standards and Schedules." 2026 edition. https://www.flsenate.gov/Laws/Statutes/2026/893.03.
14. Minnesota Office of the Revisor of Statutes. 2026. "Minnesota Statutes, Section 152.02 — Schedules of Controlled Substances." https://www.revisor.mn.gov/statutes/cite/152.02.
15. World Health Organization Collaborating Centre for Drug Statistics Methodology. 2026. "ATC/DDD Index — N06AX14 tianeptine." https://atcddd.fhi.no/atc_ddd_index/.

Stereochemistry and the enantiomers

16. Aslani, S., J. Nafie, M. F. Wahab, et al. 2025. "Tianeptine: Enantiomeric Separations, Structural Assignment, and Biological Interactions." Talanta 294: 128197. https://doi.org/10.1016/j.talanta.2025.128197.
17. Oluyomi, A. O., K. P. Datla, and G. Curzon. 1997. "Effects of the (+) and (−) Enantiomers of the Antidepressant Drug Tianeptine on 5-HTP-Induced Behaviour." Neuropharmacology 36 (3): 383–387. https://doi.org/10.1016/s0028-3908(97)00016-6.
18. Morris, R. G., S. Kelly, D. Burney, et al. 2001. "Tianeptine and Its Enantiomers: Effects on Spatial Memory in Rats with Medial Septum Lesions." Neuropharmacology 41 (2): 272–281. https://doi.org/10.1016/s0028-3908(01)00058-2.
19. Nishio, Y., C. W. Lindsley, and A. M. Bender. 2024. "Classics in Chemical Neuroscience: Tianeptine." ACS Chemical Neuroscience 15 (21): 3863–3873. https://doi.org/10.1021/acschemneuro.4c00519.

Analytical methods

20. Gaulier, J. M., P. Marquet, E. Lacassie, et al. 2000. "High-Performance Liquid Chromatographic Determination of Tianeptine in Plasma Applied to Pharmacokinetic Studies." Journal of Chromatography B 748 (2): 407–414. https://doi.org/10.1016/s0378-4347(00)00368-6.
21. Tatar Ulu, S. 2006. "Determination of Tianeptine in Human Plasma Using High-Performance Liquid Chromatography with Fluorescence Detection." Journal of Chromatography B 834 (1–2): 62–67. https://doi.org/10.1016/j.jchromb.2006.02.027.
22. Khedr, A. 2007. "High-Performance Liquid Chromatographic Stability Indicating Assay Method of Tianeptine Sodium with Simultaneous Fluorescence and UV Detection." Journal of Chromatographic Science 45 (6): 305–310. https://doi.org/10.1093/chromsci/45.6.305.
23. Abdullatef, O. A., S. Morshedy, Y. M. Khalifa, et al. 2021. "Spectrofluorimetric Determination of Tianeptine Using Its Quenching Effect on Vilazodone." Spectrochimica Acta Part A 251: 119412. https://doi.org/10.1016/j.saa.2020.119412.
24. Gazy, A. A., H. Mahgoub, E. F. Khamis, et al. 2006. "Differential Pulse, Square Wave and Adsorptive Stripping Voltammetric Quantification of Tianeptine in Tablets." Journal of Pharmaceutical and Biomedical Analysis 41 (4): 1157–1163. https://doi.org/10.1016/j.jpba.2006.02.039.
25. Boiret, M., L. Meunier, and Y. M. Ginot. 2011. "Tablet Potency of Tianeptine in Coated Tablets by Near Infrared Spectroscopy: Model Optimisation, Calibration Transfer and Confidence Intervals." Journal of Pharmaceutical and Biomedical Analysis 54 (3): 510–516. https://doi.org/10.1016/j.jpba.2010.09.029.
26. Ulu, S. T., and Z. Aydogmus. 2008. "A New Spectrophotometric Method for the Determination of Tianeptine in Tablets Using Ion-Pair Reagents." Chemical and Pharmaceutical Bulletin 56 (12): 1635–1638. https://doi.org/10.1248/cpb.56.1635.
27. Szafarz, M., A. Wencel, K. Pociecha, et al. 2018. "Pharmacokinetic Study of Tianeptine and Its Active Metabolite MC5 in Rats Following Different Routes of Administration Using a Novel Liquid Chromatography Tandem Mass Spectrometry Analytical Method." Naunyn-Schmiedeberg's Archives of Pharmacology 391 (2): 185–196. https://doi.org/10.1007/s00210-017-1448-2.

Degradation, formulation and pharmacokinetics

28. Cruz Muñoz, E., V. Termopoli, M. Orlandi, et al. 2024. "Non-Targeted Identification of Tianeptine Photodegradation Products in Water Samples by UHPLC-QTOF MS/MS." Chemosphere 361: 142534. https://doi.org/10.1016/j.chemosphere.2024.142534.
29. Lee, Y. J., and J. E. Kim. 2022. "In Vitro–In Vivo Correlation of Tianeptine Sodium Sustained-Release Dual-Layer Tablets." Molecules 27 (9): 2828. https://doi.org/10.3390/molecules27092828.
30. El-Setouhy, D. A., and N. S. Abd El-Malak. 2010. "Formulation of a Novel Tianeptine Sodium Orodispersible Film." AAPS PharmSciTech 11 (3): 1018–1025. https://doi.org/10.1208/s12249-010-9464-2.
31. Zheng, R., and B. H. Kim. 2014. "Pharmacokinetic and Bioequivalence Assessment of Two Formulations of Tianeptine Sodium in Healthy Male Volunteers." International Journal of Clinical Pharmacology and Therapeutics 52 (9): 817–823. https://doi.org/10.5414/cp202128.

Pharmacological literature

32. Gassaway, M. M., M. L. Rives, A. C. Kruegel, et al. 2014. "The Atypical Antidepressant and Neurorestorative Agent Tianeptine Is a μ-Opioid Receptor Agonist." Translational Psychiatry 4: e411. https://doi.org/10.1038/tp.2014.30.
33. Samuels, B. A., K. M. Nautiyal, A. C. Kruegel, et al. 2017. "The Behavioral Effects of the Antidepressant Tianeptine Require the Mu-Opioid Receptor." Neuropsychopharmacology 42 (10): 2052–2063. https://doi.org/10.1038/npp.2017.60.
34. Han, J., V. Andreu, C. Langreck, et al. 2022. "Mu Opioid Receptors on Hippocampal GABAergic Interneurons Are Critical for the Antidepressant Effects of Tianeptine." Neuropsychopharmacology 47 (7): 1387–1397. https://doi.org/10.1038/s41386-021-01192-2.
35. Algera, H., R. van der Schrier, D. Cavalla, et al. 2022. "Respiratory Effects of the Atypical Tricyclic Antidepressant Tianeptine in Human Models of Opioid-Induced Respiratory Depression." Anesthesiology 137 (4): 446–458. https://doi.org/10.1097/ALN.0000000000004324.
36. Baird, T. R., H. I. Akbarali, W. L. Dewey, et al. 2022. "Opioid-like Adverse Effects of Tianeptine in Male Rats and Mice." Psychopharmacology 239 (7): 2187–2199. https://doi.org/10.1007/s00213-022-06093-w.
37. Burt, S. G., G. D. Phillips, J. J. Lambert, et al. 2026. "The Atypical Antidepressant Tianeptine Causes Opioid-Receptor-Dependent Beta Oscillations in the Rat Hippocampus." Neuropharmacology 294: 110953. https://doi.org/10.1016/j.neuropharm.2026.110953.
38. McEwen, B. S., S. Chattarji, D. M. Diamond, et al. 2010. "The Neurobiological Properties of Tianeptine (Stablon): From Monoamine Hypothesis to Glutamatergic Modulation." Molecular Psychiatry 15 (3): 237–249. https://doi.org/10.1038/mp.2009.80.
39. Wagstaff, A. J., D. Ormrod, and C. M. Spencer. 2001. "Tianeptine: A Review of Its Use in Depressive Disorders." CNS Drugs 15 (3): 231–259. https://doi.org/10.2165/00023210-200115030-00006.

Forensic and toxicological literature

40. Proença, P., H. Teixeira, J. Pinheiro, et al. 2007. "Fatal Intoxication with Tianeptine (Stablon)." Forensic Science International 170 (2–3): 200–203. https://doi.org/10.1016/j.forsciint.2007.03.035.
41. Bakota, E. L., W. C. Samms, T. R. Gray, et al. 2018. "Case Reports of Fatalities Involving Tianeptine in the United States." Journal of Analytical Toxicology 42 (7): 503–509. https://doi.org/10.1093/jat/bky023.
42. El Zahran, T., J. Schier, E. Glidden, et al. 2018. "Characteristics of Tianeptine Exposures Reported to the National Poison Data System — United States, 2000–2017." MMWR. Morbidity and Mortality Weekly Report 67 (30): 815–818. https://doi.org/10.15585/mmwr.mm6730a2.
43. Marraffa, J. M., C. M. Stork, R. S. Hoffman, et al. 2018. "Poison Control Center Experience with Tianeptine: An Unregulated Pharmaceutical Product with Potential for Abuse." Clinical Toxicology 56 (11): 1155–1158. https://doi.org/10.1080/15563650.2018.1476694.
44. Quadir, M., N. I. Rine, J. Badeti, et al. 2025. "Tianeptine Exposures Reported to United States Poison Centers, 2015–2023." Journal of Medical Toxicology 21 (1): 30–41. https://doi.org/10.1007/s13181-024-01053-6.
45. Hershey, H. L., E. M. Onyango, K. Durst, et al. 2024. "Tianeptine-Involved Emergency Department Visits, Fatal Overdoses, and Substance Seizures in Tennessee, 2021–2023." Drug and Alcohol Dependence Reports 12: 100272. https://doi.org/10.1016/j.dadr.2024.100272.
46. Dempsey, S. K., J. L. Poklis, K. Sweat, et al. 2017. "Acute Toxicity from Intravenous Use of the Tricyclic Antidepressant Tianeptine." Journal of Analytical Toxicology 41 (6): 547–550. https://doi.org/10.1093/jat/bkx034.
47. Lauhan, R., A. Hsu, A. Alam, et al. 2018. "Tianeptine Abuse and Dependence: Case Report and Literature Review." Psychosomatics 59 (6): 547–553. https://doi.org/10.1016/j.psym.2018.07.006.
48. Bence, C., A. Bonord, C. Rebillard, et al. 2016. "Neonatal Abstinence Syndrome Following Tianeptine Dependence During Pregnancy." Pediatrics 137 (1): e20151414. https://doi.org/10.1542/peds.2015-1414.
49. Ikeri, K., A. Anderson, F. Eyal, et al. 2024. "Neonatal Opioid Withdrawal Syndrome Following Prenatal Use of Supplements Containing Tianeptine." Pediatrics 153 (2): e2023062382. https://doi.org/10.1542/peds.2023-062382.
50. Anand, A., M. R. Alessi, R. A. Toussi, et al. 2026. "Tianeptine Misuse and Addiction: A Systematic Review of Withdrawal, Toxicity, and Clinical Management." Drug and Alcohol Dependence 285: 113196. https://doi.org/10.1016/j.drugalcdep.2026.113196.
51. Parnia, S., L. Jain, M. Ali, et al. 2025. "Gas Station Heroin — Tianeptine and Its Impact: A Systematic Review and Exploratory Analysis." BMC Public Health 25: 3591. https://doi.org/10.1186/s12889-025-24666-0.
52. Smith, K. E., J. M. Rogers, J. C. Strickland, et al. 2021. "When an Obscurity Becomes Trend: Social-Media Descriptions of Tianeptine Use and Associated Atypical Drug Use." The American Journal of Drug and Alcohol Abuse 47 (4): 455–466. https://doi.org/10.1080/00952990.2021.1904408.
53. Counts, C. J., A. V. Spadaro, T. A. Cerbini, et al. 2025. "An Outbreak of Synthetic Cannabinoid-Adulterated Tianeptine Products in New Jersey — Case Series." Journal of Medical Toxicology 21 (2): 253–259. https://doi.org/10.1007/s13181-025-01068-7.
54. Rouby, F., V. Pradel, E. Frauger, et al. 2012. "Assessment of Abuse of Tianeptine from a Reimbursement Database Using 'Doctor-Shopping' as an Indicator." Fundamental & Clinical Pharmacology 26 (2): 286–294. https://doi.org/10.1111/j.1472-8206.2010.00906.x.
55. Edinoff, A. N., S. Sall, S. P. Beckman, et al. 2023. "Tianeptine, an Antidepressant with Opioid Agonist Effects: Pharmacology and Abuse Potential, a Narrative Review." Pain and Therapy 12 (5): 1121–1134. https://doi.org/10.1007/s40122-023-00539-5.