MDMA
Two positive Phase 3 trials, a 10–1 advisory vote against, and the rejection that reshaped a field.
MDMA is the compound every other program in this series is positioned against. Synthesized by Merck in 1912 as a synthetic intermediate, rediscovered by Alexander Shulgin in 1976, used quietly by psychotherapists for a decade, criminalized in 1985 over an administrative law judge's objection, and then carried through two successful Phase 3 trials by a non-profit advocacy organization — before the FDA declined to approve it in August 2024. The rejection was not about whether patients improved. It was about trial conduct, functional unblinding and data integrity, and it is the single most consequential regulatory event in modern psychedelic medicine.
It was a synthetic intermediate, not a diet pill.
Merck made it in 1912 en route to a hemostatic agent. The widely repeated “developed as an appetite suppressant” story is false, and its persistence is a lesson in how compound origin myths form.
Its own judge recommended against Schedule I.
DEA's administrative law judge, Francis Young, recommended Schedule III on the strength of the psychotherapeutic evidence. The Administrator overrode him. The scheduling was contested from the inside.
Mechanism: transporter release, plus oxytocin.
A substrate-type releaser with marked SERT preference, driving serotonin, oxytocin and vasopressin release. This is the same transporter mechanism as methylone — with a VMAT2 component methylone largely lacks.
The neurotoxicity literature contains a retracted landmark.
Ricaurte's 2002 Science paper reporting dopaminergic neurotoxicity was retracted in 2003 after the animals turned out to have received methamphetamine. Serotonergic concerns remain live; that one does not.
Both Phase 3 trials worked, at large effect sizes.
MAPP1: d = 0.91, 67% no longer met PTSD criteria. MAPP2: d = 0.70, 71%. These remain the largest reported PTSD treatment effects in the modern literature — and they were not enough.
The rejection was about method, not molecule.
Functional unblinding, inseparable psychotherapy, trial-conduct violations, incomplete cardiac data, unproven durability. Every subsequent program in this field is designed around that list.
A century, in eleven moments
Origins:
a step on the way to something else
MDMA was first prepared at Merck in Darmstadt in 1912 by the chemist Anton Köllisch, working on a program aimed at hemostatic substances. It was an intermediate — a step in a synthetic route toward hydrastinine analogues — and Merck's patent application, filed on 24 December 1912, claims the process rather than any therapeutic use of the compound itself.1
MDMA is routinely described as having been “developed by Merck as an appetite suppressant.” Merck's own archival review found no evidence for this; the compound was never developed as an anorectic and never tested in humans by the company. The story appears to have been reconstructed backwards from MDMA's amphetamine skeleton.1 The parallel with methylone is exact — a compound's stated patent context becomes, in retelling, a development program that never existed. Treat all “originally developed as X” claims in this field as requiring a primary source.
Merck revisited the compound sporadically over the following decades without pursuing it. In 1953–54 the US Army Chemical Center ran animal toxicity studies on a series of compounds including MDMA, under code EA-1475; the work was declassified in 1969 and is sometimes overstated as an interrogation program when it was a toxicity screen.2
The compound became psychoactively known to the world through Alexander Shulgin, who resynthesised it in 1976 after a student at San Francisco State University described its effects to him. Shulgin self-experimented, was struck by the result, and — critically — introduced it to the psychologist Leo Zeff, who came out of retirement to teach its use to other therapists. In 1978 Shulgin and David Nichols published the first account of its human pharmacology, describing an easily controlled altered state with emotional and sensual overtones.3
Note the structure of that history, because it recurs across this whole series: Shulgin appears again as the human characteriser of a compound made decades earlier for unrelated reasons. He performed the same role for methylone, for 4-OH-DiPT and for dozens of others. The modern psychedelic pipeline is, to a striking degree, the industrialisation of one chemist's notebooks.
Adam, Ecstasy, and the 1985 scheduling fight
Between roughly 1977 and 1985 MDMA occupied a position no other compound in this series has held: it was legal, it was not a mass-market drug, and it was in routine clinical use by an informal network of psychotherapists who called it Adam. Estimates of how many therapists used it vary widely and none are reliable, but the practice was real, deliberately low-profile, and generated a body of clinical impression rather than published trials.4
That quiet arrangement ended because of distribution, not because of harm. In the early 1980s MDMA was rebranded Ecstasy and sold openly — most conspicuously over the counter in Dallas and Austin bars, purchasable by credit card. The scale and visibility of that trade made federal action inevitable. On 1 July 1985 DEA placed MDMA in Schedule I on an emergency basis.4,5
What followed is the part most often omitted. Researchers and therapists requested a hearing, and DEA's own administrative law judge, Francis L. Young, heard testimony over several months. Young concluded that MDMA had an accepted medical use and recommended Schedule III. The DEA Administrator rejected the recommendation and scheduled it in Schedule I. A 1987 First Circuit decision vacated that placement on a technical ground; DEA promptly re-scheduled, and the Schedule I designation became final in 1988.5
MDMA is the only compound in this series whose Schedule I placement was formally recommended against by the scheduling agency's own judge — and imposed anyway.
The cultural history after 1985 is the familiar one: UK acid house and the 1988 “Second Summer of Love,” the global rave economy, and a public-health discourse organized around a small number of highly publicised deaths. The pharmacological reality behind those deaths is mostly hyperthermia and hyponatraemia — thermoregulatory failure in hot, crowded settings, and water intoxication from over-correction — rather than direct dose-dependent toxicity. Adulteration compounded it; as the methylone report documents, a substantial share of what was sold as MDMA in the 2010s was not MDMA.
Against that backdrop, Rick Doblin founded MAPS in 1986 with the explicit aim of making MDMA a licensed medicine. That a single advocacy non-profit carried a Schedule I compound to a filed NDA over thirty-eight years is a genuine achievement. As Section VII argues, it was also part of the problem.
Mechanism of action
MDMA is a substrate-type releaser at the plasma-membrane monoamine transporters. It is taken up by SERT, DAT and NET, induces a conformational change that reverses their normal direction, and drives non-exocytotic efflux of cytoplasmic monoamine — with a marked preference for serotonin.6 It also disrupts vesicular storage through VMAT2, which is the step that distinguishes it pharmacologically from methylone.
Three features account for the entactogenic profile that distinguishes MDMA from a stimulant:
- Serotonin dominance. The SERT-weighted release profile produces mood elevation and emotional openness rather than the dopamine-weighted drive of amphetamine or cocaine.
- Neurohormonal release. MDMA raises plasma oxytocin, vasopressin and cortisol. The oxytocin component is the leading candidate mechanism for the prosocial, trust-enhancing quality that made the compound attractive to psychotherapists in the first place.6
- Fear memory effects. MDMA modulates fear memory reconsolidation, enhances fear extinction, and promotes openness and prosocial behavior — the mechanistic rationale cited in the MAPP2 paper itself for its use in PTSD.10
The third point deserves emphasis, because it is the thread running through this entire series. Fear extinction enhancement is the shared translational endpoint of MDMA, methylone and 4-OH-DiPT — a β-keto amphetamine, a 4-hydroxytryptamine and MDMA itself, arriving at the same behavioral mechanism by transporter release, transporter release without VMAT2, and 5-HT2A agonism respectively. Either extinction learning is a genuine common path to trauma- and anxiety-spectrum benefit, or it is the assay this field knows how to run.
Unlike the classic psychedelics, MDMA is not primarily a 5-HT2A agonist and is not hallucinogenic at therapeutic doses. It occupies a distinct pharmacological category — the entactogens — and shares that category with methylone, which is why the two are properly compared and why psilocybin and LSD are not the right comparators for either.
Two entactogens, one mechanism, different consequences
MDMA against methylone — the only genuine like-for-like comparison in this series, and the reason TSND-201 exists.
The strategic read. TSND-201 is, in effect, MDMA with the two properties that caused FDA trouble engineered out: the psychedelic-adjacent subjective intensity that made blinding impossible, and the elaborate psychotherapy that made the drug effect inseparable from the treatment package. Whether that is a genuine improvement or merely a smaller version of the same effect is exactly what methylone's Phase 3 will test.
Neurotoxicity, and a retraction that still distorts the debate
MDMA's neurotoxicity literature is large, genuinely contested, and contains one of the most consequential retractions in modern neuroscience. Separating the two is necessary before any risk statement can be made.
The real signal. Repeated high-dose MDMA produces long-lasting reductions in serotonergic markers — 5-HT content, SERT binding and tryptophan hydroxylase activity — in rodents and non-human primates. Whether this constitutes axonal degeneration or reversible down-regulation has been argued for three decades. Human evidence, largely from heavy recreational users, suggests reduced SERT binding and some cognitive deficits, but is confounded by polydrug use, adulteration, dose uncertainty and pre-existing differences. Hyperthermia strongly potentiates the effect in animals, which is why ambient temperature matters so much to interpretation.7
In 2002 George Ricaurte's group published in Science a report that MDMA caused severe dopaminergic neurotoxicity in non-human primates, implying a risk of parkinsonism from recreational doses. The finding was widely covered and cited in policy debate. In 2003 the paper was retracted: the animals had been dosed with methamphetamine, not MDMA, following a labeling error at the supplier.8
Two lessons, pulling in opposite directions. The dopaminergic-parkinsonism claim has no evidentiary standing and should never be repeated. But the episode also hardened advocacy positions in a way that has made sober discussion of the serotonergic risk — which is not retracted and is not resolved — harder than it should be. A reader encountering strong claims in either direction should check whether the source distinguishes the two.
Acute risks are better characterized and clinically more relevant: hyperthermia, hyponatraemia from excess water intake, serotonin syndrome in combination with serotonergic agents, hepatotoxicity, and cardiovascular strain. Sympathomimetic increases in blood pressure and heart rate are expected. In MAPP2, cardiac treatment-emergent adverse events were low in rate, and blood-pressure and pulse increases were mild and transient.10 FDA nonetheless considered the cardiac safety database incomplete — a reminder that “low rate in 104 patients” is not a cardiovascular safety package.
Abuse liability is genuine, and unlike the classic psychedelics MDMA is self-administered by animals. It carries real, if moderate, dependence potential — a consideration that applies equally to methylone and does not apply to psilocybin or LSD. Any approval would require an abuse-liability and diversion package that the tryptamine programs largely escape.
MAPP1 and MAPP2: the trials worked
It is important to be clear about this before discussing the rejection: both pivotal trials met their primary endpoints, at effect sizes larger than anything else in the PTSD literature, and both were published in Nature Medicine.
The treatment package was three monthly eight-hour dosing sessions with a flexible dose of 80–120 mg plus an optional supplemental half-dose, delivered by a two-therapist team, wrapped in non-drug preparatory and integrative psychotherapy. The comparator was inactive placebo with the same therapy. The primary outcome was CAPS-5 total severity, rated by a blinded centralised independent rater pool.9,10
MAPP1 (Mitchell et al. 2021; n = 90; severe PTSD; mean baseline CAPS-5 44.1; mean duration 14.1 years) reported an LS-mean CAPS-5 change of −24.5 versus −12.6 for placebo-with-therapy, between-group d = 0.91. At endpoint, 67% no longer met PTSD diagnostic criteria versus 32% on placebo, and 33% met remission criteria versus 5%.9,11
MAPP2 (Mitchell et al. 2023; n = 104; moderate-to-severe PTSD; mean baseline CAPS-5 ≈ 39; mean duration 16.2 years; a more diverse cohort than MAPP1) reported d = 0.70, with 71% no longer meeting criteria versus 48%, and 46% remission versus 21%. Dropout was 1.9% in the MDMA arm against 15.7% on placebo-with-therapy — remarkable against the 46–56% dropout reported for prolonged exposure and cognitive processing therapy in veterans.10,11
Australia acted on this evidence base before the US did. From 1 July 2023 the Therapeutic Goods Administration permitted authorised psychiatrists to prescribe MDMA for PTSD — making Australia the only country where MDMA can be legally prescribed, a status it still held in 2026.12
MAPP1 and MAPP2 outcomes
Loss of PTSD diagnosis and remission, MDMA-assisted therapy versus placebo-with-therapy. Both trials met their primary endpoint; note how well the placebo arms performed.
The gap narrows in the better trial. MAPP2 enrolled a broader, more diverse, less severe cohort — and the effect size fell from 0.91 to 0.70 while the placebo response rose from 32% to 48%. That is the expected direction as a trial becomes more representative, and it is the same pattern seen in psilocybin's move from academic Phase 2 to industry Phase 3. Read every large psychedelic Phase 2 effect with this trajectory in mind.
August 2024: what actually happened
FDA accepted Lykos Therapeutics' NDA for midomafetamine capsules on 9 February 2024 with priority review. On 4 June 2024 the Psychopharmacologic Drugs Advisory Committee voted heavily against — 9–2 that the data did not show the treatment was effective, and 10–1 that the benefits did not outweigh the risks. On 9 August 2024 FDA issued a Complete Response Letter and requested an additional Phase 3 study.13,14
The CRL was published in September 2025, when FDA released eighty-nine previously unpublished response letters as part of a shift toward disclosure.15 It is worth reading precisely, because the popular framing — that FDA rejected a psychedelic on ideological grounds — is not what the record shows. The stated deficiencies were:
The week of the CRL also brought the retraction of three papers in Psychopharmacology reporting data from an earlier MAPS study, on the grounds of unethical conduct at a trial site that had not been properly disclosed.16 Whatever one concludes about the efficacy question, the data-integrity concerns were not manufactured.
Lykos reorganised on 15 August 2024 — a ~75% workforce reduction, Rick Doblin's departure from the board, and the appointment of former Janssen executive David Hough (who had worked on the esketamine approval) to lead regulatory engagement.17 As of mid-2026 no resubmission date has been announced, and a new Phase 3 would take years.18
MAPS was founded to make MDMA a medicine. That mission produced the trials — and it also meant the sponsor, many investigators, and a substantial fraction of participants shared a prior commitment to the outcome. In a trial where blinding fails and the primary endpoint is a clinician-rated interview, that is not a reputational problem, it is a measurement problem. The lesson every subsequent program drew is visible in this series: corporate sponsors, low-dose active comparators, monitoring rather than manualised therapy, registrational endpoints, and no advocacy.
The commercial landscape MDMA reshaped
Every psychedelic or psychedelic-adjacent compound in active late-stage psychiatric development. MDMA is the only entry that reached an FDA decision — and the only one that failed.
4-PO-DMT
same molecule
CYB003 / HLP003
lysergide D-tartrate
βk-MDMA
via luvesilocin
mebufotenin benzoate
mebufotenin
N,N-dimethyltryptamine
CYB004
midomafetamine
single enantiomer
How to read this table. Two structural facts stand out. First, almost nobody owns their molecule — psilocybin, LSD, DMT, 5-MeO-DMT, MDMA, methylone and 4-OH-DiPT are all unpatentable as compositions, so every program competes on formulation, route, deuteration or prodrug chemistry. Second, the field has converged on shortening the session: intranasal, inhaled, buccal, subcutaneous and deuterated routes all exist to cut monitored clinic hours, the binding constraint on this entire class.
MDMA's real legacy is in the design column. Look at what everyone else is doing: low-dose active comparators rather than inert placebo (Compass, Reunion, MindMed's 50 µg arm), monitoring rather than manualised psychotherapy (Transcend, GH Research), adjunctive designs that avoid antidepressant washout (Cybin), and corporate sponsors with no advocacy history. Every one of those choices is a direct response to the five deficiencies in the August 2024 CRL. MDMA failed, and in failing it wrote the protocol template for the entire field.
Established, probable, speculative
Grading is ours. A — multiple independent sources, consistent. B — solid primary evidence, limited replication. C — single source, sponsor-derived, or contested.
Established · A
- It is a substrate-type monoamine releaser with SERT preference, and releases oxytocin.6
- Merck 1912, as a synthetic intermediate — not as an appetite suppressant.1
- Shulgin resynthesised it in 1976; the first human account published 1978.3
- Emergency Schedule I 1985 over the ALJ's Schedule III recommendation; permanent 1988.5
- Both Phase 3 trials met their primary endpoint and published in Nature Medicine.9,10
- FDA issued a CRL on 9 Aug 2024 and required an additional Phase 3.14,15
Probable · B
- MDMA-assisted therapy substantially reduces PTSD symptoms versus placebo-with-therapy.9,10
- Dropout is markedly lower than for standard trauma-focused psychotherapy.10
- Repeated high-dose exposure produces lasting serotonergic marker reductions in animals.7
- It enhances fear extinction and modulates memory reconsolidation.10
- Acute risk is dominated by hyperthermia and hyponatraemia, not direct dose toxicity.7
Speculative · C
- How much of the Phase 3 effect is drug rather than expectancy. This is the whole question, and it is open.14
- Whether the psychotherapy component is necessary, sufficient, or separable.
- Durability beyond two months from three sessions.10
- Clinically meaningful serotonergic neurotoxicity at therapeutic doses in humans.7
- Whether and when a resubmission is viable.18
Research gaps, in order of consequence
Nothing in the record establishes what MDMA does for PTSD absent an eight-hour two-therapist session. Until a design separates the two, the CRL's central objection stands unanswered.
With ~40% MDMA-experienced participants and unblindable effects, the field still lacks a validated instrument to partition drug effect from expectancy. Every program in Figure 4 needs this and none has it.
A trial enrolling only MDMA-naive participants would speak directly to both generalisability and expectancy. It has not been done, and it is the most informative single study anyone could run.
Two months is not durability for a chronic condition. Nor is there systematic data on post-trial MDMA use — a question the MAPP2 authors themselves flag as needing long-term study.
A sympathomimetic in a population with elevated cardiometabolic risk needs more than 194 pooled Phase 3 patients. FDA said so explicitly.
Australia has been prescribing since July 2023. Systematic outcome and safety data from that real-world population would be the most valuable evidence available anywhere — and it does not appear to have been published.
Sources
P1 peer-reviewed primary research or official instrument · P2 review, book or historical scholarship · P3 sponsor communication or trade press · P4 general reference, non-load-bearing.