Ecstasy for PTSD: Not Ready for Prime Time

Hundreds of desperate combat veterans with Post-Traumatic Stress Disorder (PTSD) are reportedly seeking experimental treatment with an illegal drug from a husband-wife team in South Carolina. The Bonhoefers recently published a study showing that adding MDMA (ecstasy, the party drug) to psychotherapy was effective in eliminating or greatly reducing the symptoms of refractory PTSD. It was widely covered in the media, for instance in this article in the NY Times. It was only a small preliminary study, and the treatment is not yet ready for prime time; but media reports have sparked enthusiasm not justified by the evidence.

The Study

The study is available online. To summarize the highlights:

  • Subjects were 20 patients with chronic PTSD refractory to both psychotherapy and psychopharmacology.  85% were women, mostly victims of sexual assault or childhood abuse, and there was only 1 subject with combat stress.
  • Randomized.
  • Placebo controlled: 12 patients got the active drug, 8 got an inactive placebo
  • Method: preliminary and follow-up psychotherapy, 2 day-long sessions with 2 co-therapists and an overnight hospital stay. Subjects got two doses of MDMA during each session and alternated conversation with periods where they were encouraged to focus on introspection.
  • Assessment instruments were CAPS, a structured interview, and IES-R, a self-report of response to stress. Neurocognitive testing was done and vital signs were monitored.
  • Rescue medication with Zolpidem and benzodiazepines was allowed when needed. This did not appear to influence results.
  • Protocol amendment: After safety and efficacy were established, an amendment to the protocol was approved so that the last 9 subjects received a third dose of MDMA to prolong the effect and gather data about dosage for future trials.
  • Crossover extension: 7 out of 8 subjects from the placebo group were enrolled in a subsequent open-label phase with MDMA; one declined because she had already improved so much on the placebo.
  • Exit poll: 95% correctly guessed which group they were assigned to.


  • Clinical response was defined as >30% reduction from baseline in CAPS total severity score. The clinical response was 83% in the MDMA group compared to 25% in the placebo group.
  • 100% of the 7 subjects from the placebo group who were subsequently given MDMA open-label had a clinical response.
  • All 3 subjects who had been unable to work due to PSTD were able to return to work.
  • There were no significant changes in neurocognitive tests.
  • Side effects occurred, ranging from jaw tightness to insomnia, resolving over hours or days, sometime requiring short-term drug treatment; none were felt to be serious.

Limitations of the study

This was only a small pilot study, not the kind of evidence needed to justify forging ahead with treatment of patients with PTSD. The subjects were almost all women who were rape victims. Its applicability to war trauma is entirely unknown. Further studies with combat veterans are in the works.

The study is characterized as blinded; but the blinding didn’t work, since 95% of subjects were able to tell which group they were in.

Subjects may have been influenced by the reputation of MDMA as a recreational drug that produces euphoria. MDMA is illegal, and this study required a special Investigational New Drug (IND) permit. It was financed by the Multidisciplinary Association for Psychedelic Studies.


In a 2010 news report, Mithoefer

expressed concern that publicity about the study might lead some people to self-medicate with ecstasy. PTSD patients should not use the drug on their own, Mithoefer said, because there are risks, such as elevated blood pressure and pulse. Also, ecstasy purchased on the street may not be pure MDMA — or even contain no MDMA at all.

In the NY Times article there is a quotation from Brig. Gen. Loree Sutton, a psychiatrist who recently retired from the Army:

When it comes to the health and well-being of those who serve, we should leave our politics at the door and not be afraid to follow the data.  There’s now an evidence base for this MDMA therapy and a plausible story about what may be going on in the brain to account for the effects.

That’s inaccurate; there is not an “evidence base.” There is only a pilot study with scanty, preliminary, unconfirmed evidence.

A military spokesman put the cart before the horse:

Completing the studies necessary to make this treatment available will require increasing financial and political support from both within and outside the military,” he said. “We provide men and women in the armed forces with the most advanced tools of war. It’s time we gave them the most advanced tools of healing, too.”

It’s not a matter of doing studies necessary to make the treatment available, but rather a matter of doing the studies necessary to determine whether the treatment really is effective and is a feasible option.

Other Questions

The use of another illegal drug, LSD, has been advocated by some psychiatrists for a multitude of uses. Experiments using illegal drugs are always problematic, because researchers tend to be advocates for legalization and may be biased by their emotional investment.

Prolonged exposure therapy is a widely recognized therapy for PTSD. The rationale for using MDMA is to suppress anxiety and allow patients to talk about their traumatic experiences without feeling excessive fear and hyperarousal. Various possible (and plausible) mechanisms for its usefulness are discussed in the introduction to the study. Another technique that has been used to facilitate patients’ engaging with psychotherapy is Eye Movement Desensitization and Reprocessing (EMDR), where the eye movements themselves don’t actually do anything.

The patients in the placebo group improved, too; so a large part of the benefit could be attributed to the psychotherapy.  Should the specific psychotherapy techniques they used be further investigated and improved? Was there something special about the effects of MDMA, or might legal anti-anxiety drugs be just as effective?

After WWII, soldiers suffering from “combat fatigue” were given sodium pentothal (“truth serum”) and were allegedly cured by re-experiencing the trauma.  Such treatments apparently didn’t really work that well, since they quickly fell out of favor. Most references are from the 1940’s. One flight surgeon found that distributing liquor at mission debriefings was effective at “settling” his fliers, but I don’t think that was ever formally studied.

Is it essential that patients be guided to re-live their traumatic experience? It sounds plausible, but I don’t know and doubt that anyone does. Is it possible that PTSD could be treated without that, by accentuating the positive and concentrating on avoidance and coping skills? I don’t know. Is it possible that PTSD could be prevented by early debriefings right after the event with the assistance of alcoholic beverages? I don’t know, but it’s an intriguing thought. Maybe some research group could get a grant from Jack Daniel’s.


The addition of MDMA to psychotherapy might prove helpful in refractory cases of PTSD, but the preliminary results of this one small pilot study will need to be confirmed by larger studies in combat veterans before the treatment can recommended to patients. Unfortunately, the history of medicine is full of equally promising treatments that didn’t pan out. Probably the most typical course is this: a strongly positive pilot study is followed by larger studies with weaker positive results, then by studies with equivocal or negative results, then reports of serious side effects or other problems, and eventually by rejection of the treatment. I sincerely hope this one will prove an exception to that course, for the sake of suffering veterans with few remaining options. But I am not optimistic.



Posted in: Clinical Trials, Neuroscience/Mental Health

Leave a Comment (24) ↓

24 thoughts on “Ecstasy for PTSD: Not Ready for Prime Time

  1. geo says:

    There was a recent programme on Channel 4 which had volunteers taking MDMA as part of an exciting live experiment!

    This fun TV involved the work of respected researchers like David Nutt, but I’ve forgotten most of the results they discussed. The participant who had the worst reaction was a former soldier, who felt that some of his training for resisting interrogation (or perhaps his own personality and preferences) led to him resisting the affects of the drug.

    “Experiments using illegal drugs are always problematic, because researchers tend to be advocates for legalization and may be biased by their emotional investment.”

    I get this sense too, but I also think that these sorts of biases tend to be fairly common in those testing any treatment, and there will be many with prejudices against the value of drugs used recreationally. Whether they acknowledge it or not, I think that most researchers investigating a treatment start with a desire to show efficacy, or else debunk a treatment they feel has been over-promoted.

  2. doc_leigh says:

    i have yet to read the long-term follow up part of the study by Mithoefer et al ( in fact, i just found it… i had read through the *new* MDMA for PTSD study published at the same time, which is even weaker than the older study you are describing here. you can find the new paper here:

    the new study used an active placebo, which was effective at improving the blinding. of course, with improved experimenter blinding, the observed therapeutic effect diminished substantially. by substantially, i mean to the point where it was not statistically significant. not to mention the fantastically poor study design that was simply not set up to observe long-term effects, the fact that there was NO actual placebo control whatsoever, and that their sample size in some treatment groups was too small to even run inferential statistics.

    of course, the study with no significant effect that was published at the same time gets no press…

  3. WilliamLawrenceUtridge says:

    I know in studies of alcohol use mostly non-alcoholic beverages as a placebo (a thin layer of alcohol is floating on top of water or some other substance) and I think that’s seen as a fairly effective placebo because most subjects genuinely think they are getting real booze. But I still wonder about such an inert placebo (lactose). The effects of MDMA are pretty apparent to the user, making me wonder how much the placebo and nocebo effects influenced the outcome. Is it really a fair test? I kinda doubt it.

  4. MattM says:

    They are doing it wrong …

  5. daedalus2u says:

    I remember reading some articles about opiates given during rescue and early treatment for casualties of war trauma and that this seemed to have an effect on the later acquisition of PTSD (opiates reduced it).

    One way that the Vietnam conflict was different than the wars in Iraq and Afghanistan is the lack of drug testing in the Vietnam Era and (as I remember from news accounts at the time), very high availability of alcohol, opiates and marijuana in theater. Was that drug use self-medication? and did that prevent PTSD and suicide in Vietnam era active duty personnel and veterans?

    PTSD from war trauma (of multiple different types) is likely to be different than PTSD from sexual assault. The time scale for their occurrence is very different. It is very likely that the time scale for their resolution would be different too.

    I am concerned that in their zeal to “fix” the problem of military PTSD and suicide, that non-rigorous or “band-aid”-type “magic bullets” are going to get all the funding and that more science based research and treatments won’t be funded.

  6. galaoxides says:

    The quote attributed to the military is actually Burge’s, the MAPS spokesman.

  7. Robb says:

    There’s also a long term follow up to this study that was just published online this month:
    This study also isn’t the only one of its kind – but the others have also been small in size.
    As some of you probably know, MDMA did not start out as a party drug but was originally used by psychotherapists in treating anxiety and other problems before it was classed Schedule I. Considering none of the trials have shown any safety issues with pure MDMA and that PTSD is on the rise, it would be nice if larger trials can expand on these results. I also don’t see how anyone could be unsure about whether they got the active or control treatment though…

  8. Quill says:

    I agree that since MDMA’s effects are so specific and powerful, it may not be possible to do a proper RCT of it. It’s the inverse of homeopathy where the placebo and the alleged substance are actually the same thing (water). In this case, as with so may other things that are intensely psychoactive, it seems beyond current techniques to have any kind of placebo.

  9. Shelley says:

    Yeah, I’m with daedalus2u on this one. I suspect PTSD from combat is very different, and there are so many problems with this small study, it is difficult to know where to begin. One certainly can’t draw any serious conclusions from it at this point.

    Alcohol has always been a form of self medication for PTSD – and it doesn’t work long term and has a multitude of very negative side effects. When I have worked with combat veterans, good old exposure therapy (without chronic use of benzodiazepines) seems most effective. I doubt that we will find a simple pill solution for a complex and multifaceted disorder like PTSD (as nice as that would be).

  10. Narad says:

    As some of you probably know, MDMA did not start out as a party drug but was originally used by psychotherapists in treating anxiety and other problems before it was classed Schedule I.

    I would quibble with the “originally.” My understanding is that Leo Zeff picked up on it (enthusiastically) as a therapeutic agent by way of Shulgin, whose use seems to have been frankly recreational at the outset.

    As for the placebo issue, Oehen et al. appear to have used 25 mg + 12.5 mg booster at 2.5 hr versus a full dose of 125 mg + 62.5 mg booster at 2.5 hr. While I’ve never tried the stuff, I’ve read nothing to suggest that it’s “intensely psychoactive.” The effects are perhaps readily apparent to someone who specializes in dosing people with it, particularly given the all-day-long psychotherapy design, but I’m wondering whether some MDMA homologue at closer to full strength might have made more sense.

  11. Robb says:

    The point about original use was more that it was originally used in a therapeutic setting with therapeutic goals rather than “just to get high and dance”. Shulgin did have more of a tinkering, exploratory, hobbyist approach but if I remember right he was horrified at the way it ultimately made its way into casual use. It’s similar in structure to amphetamines so there is no mistaking a psychoactive effect – a placebo would need to also have a subjectively noticeable stimulating effect of some kind. It also causes pupil dilation so ideally a placebo would need to do this as well otherwise it would be clear who got it and who didn’t. I don’t think a smaller “booster” would be very effective as it is also known to lead to rapid tolerance/diminished effects.

  12. Jacob V says:

    A couple of the commenter’s above seem to be implying that childhood sex abuse as a onetime incident or of limited duration. It can be, however I can assure you that many of the victims of childhood sex abuse were chronically abused over many years; and given the concurrent issues of personality development and fractured parental attachments the long term effects of childhood sex abuse, including PTSD, can last a lifetime.

    Then again I suppose I could just go with Richard Dawkins on this one and say that any child who was taught there was a hell was undoubtedly more damaged than a combat veteran.

  13. LMA says:

    While these kinds of studies may sound humorous to the average person, it is important to note as Robb does that these medications were originally developed with an eye towards helping mentally ill people. And that as Jacob V indicates, the specifics of a given trauma don’t really matter; it’s that the person has been traumatized.

    Anyway, right now the NIMH has an open and ongoing clinical trial on the effect “Special K” (ketamine) may have on major depressive episodes. The problem with these studies as I see it, and I speak as someone desperate for a breakthrough (I have bipolar disorder and have proved to be resistant to at least a dozen antidepressants and a half dozen more anti-psychotics), is the whole nature of a double blind study. I can’t offer myself up for the ketamine study for example, because in order to do the study I’d have to stop taking everything I’m on now. Mind you, I’m still depressed on the existing drugs, but I’m not suicidal like I was earlier in the year, and so to risk going off existing medications to try something new is to risk feeling like killing myself again. Even I’m not “crazy enough” to do that.

    OTOH, medications for mental illness are inherently different from those for say, heart disease. A patient can’t tell you if a new statin is doing anything for her heart, but she *can* tell you if a new medication leaves her feeling less suicidal than she did before. Long-term effects can only be monitored by continued study, but I can say with absolute certainty for example, that within a week of first taking Prozac, I felt completely different than I had in my first 21 years of life. Unfortunately, I had to take higher and higher doses (there’s a lot of similarity between street drugs and psych meds) to get that effect, and eventually I had to try other things because it didn’t work at all, but short term, it was astonishing.

    Why can’t a new type of study protocol be designed for these sorts of medications, whether they are for PTSD or depression or anxiety? A study that would recognize that double-blinding doesn’t work or is too dangerous and instead takes into account the patient’s on-going opinions instead of playing potentially fatal games by offering something-that-might-help-and-I-know-I’m-on-something and a placebo-that-can’t-help-and-I-can-feel-that-too? Why are patients always being treated as an inanimate object with the only results respected being those that can be chemically measured?

  14. mousethatroared says:

    Sorry if this is redundant, I’m afraid I don’t have time to read all the comments. But I wanted to make a comment on this

    “expressed concern that publicity about the study might lead some people to self-medicate with ecstasy. PTSD patients should not use the drug on their own, Mithoefer said, because there are risks, such as elevated blood pressure and pulse. Also, ecstasy purchased on the street may not be pure MDMA — or even contain no MDMA at all.”

    Yes! Don’t try this at home. – I was never into the recreational pills, but in my younger years I did witness some bad ecstasy experiences at social events that I would not put anyone with PTSD or anyone with a history of being sexually abused through.

    On the other hand, a while ago I heard about some research that indicated that retelling of the traumatic experience while the patient was given medication that eliminated the stress response (I’m not sure what drug, maybe benzodiazepines) was beneficial to people with PTSD, while retelling the experience without stress reduction tended to be not beneficial. I don’t know how good the research was. I think I heard about it on RadioLab, which isn’t the most reliable source for medial research interpretation. ;) If I get a chance I’ll look it up.

  15. Harriet Hall says:


    “Why are patients always being treated as an inanimate object with the only results respected being those that can be chemically measured?”

    That’s not a fair characterization. Patients are not treated as inanimate objects, but as unreliable witnesses, subject to placebo effects and to misperceptions and misattributions of their experiences.

  16. Robb says:

    You bring up some interesting points. I’m not a health professional myself but my thoughts on some of your questions are this: there are a variety of different study designs to choose from and not all of them are double blind, placebo controlled. Others could be measured against a current existing treatment for example. I don’t think placebo controlled trials mean to treat patients as inanimate objects, but the patient is not the focus of the study, the drug/treatment is. The patients are really participating to help find out if the drug/treatment works and can be used to help other people – the primary purpose isn’t for their own treatment during the trial, although some may experience benefits anyway. I guess you could look at the design as a necessary evil in order to really know if you have something that can potentially help a much larger group of people in the population at large.

    With mental health it can be trickier, as the treatment can consist of multiple components (MDMA + psychotherapy for example) and there is more of a subjective element. Scoring tests are of course used to try to quantify subjective mental health, but I think they are still far removed from objective measures like blood pressure, etc.

    So in your case, I’d imagine there would be strong ethical issues (let alone your own concerns) with your going off medication in order to participate. I’m not really sure how the ethical issues are navigated in general for studies with depressed patients potentially receiving placebo and going downhill though. I’d imagine it’s a balancing act of respecting that you are treating real people with the fact that certain design factors are needed to ensure you are getting the most accurate results, with the implication that the more accurate the results, the better you will know how well future people can benefit (or not).

  17. Agrippina says:

    Thanks for the post Dr. Hall.

    A persistent problem with media reporting on this trial since 2004 has been that anyone who raises concerns is stigmatized as a mindless anti-drug zealot. This trial set, to my knowledge, the record for IRB shopping. It’s my understanding that Copernicus IRB, the commercial IRB that brought us the Ketek protocol 3014 fiasco, approved the protocol after 8 IRBs rejected it.

    The first IRB (Western IRB) to approve the trial had to rescind approval upon further consideration. Thanks to the sponsor’s outrage at FDA regulations for the protection of research subjects, many of these documents are online:

    As for Dr Mithoefer’s concern about self-medication and the purity of street-grade E, perhaps this same concern could have been extended to research subjects in a MAPS-sponsored trial in Spain testing E for PTSD in rape victims. The trial was terminated when somebody noticed the (sole) physician quit leaving a psych grad student in charge. The source of the “study drug” was the local police impound, which is stupid beyond belief and grossly unethical. IIRC proponents complained the trial was terminated for political reasons.

    I have every sympathy for people with treatment refractory conditions but exquisite care must taken in the oversight of research conducted by ideologues. I would also prefer to see sponsors select qualified investigators with relevant experience in conducting clinical trials.

  18. Narad says:

    Shulgin did have more of a tinkering, exploratory, hobbyist approach but if I remember right he was horrified at the way it ultimately made its way into casual use.

    This bit is usually associated with the DOM story.

  19. Narad says:

    OK, sorry, I just had company leave and I’m catching up.

    I don’t think a smaller “booster” would be very effective as it is also known to lead to rapid tolerance/diminished effects.

    I suspect that you’re thinking of LSD, which is not usefully supplemented outside a narrow window. (Psilocin is a bit more forgiving in this regard. Just sayin’.) MDMA is well known as being boostable to stretch out the effect, which is what one would naively expect from an amphetamine backbone. These guys seem to have made the pharmacokinetics into something of a cottage industry.

  20. Narad says:

    While these kinds of studies may sound humorous to the average person, it is important to note as Robb does that these medications were originally developed with an eye towards helping mentally ill people.

    This is not true with respect to either MDMA or ketamine, although, speaking as someone who has had poor luck in the SSRI/SNRI/atypicals game, I can very much empathize with interest in the latter.

  21. Robb says:

    Ketamine was originally used as an anaesthetic for humans and in veterinary practice and MDMA was originally synthesized in an attempt to find an analogue to hydrastinine as part of pharmaceutical companies competing with each other. MDMA wasn’t ultimately used for anything until psychotherapists started using it in the 1970s-80s.

  22. Xplodyncow says:

    LMA, Narad, I can relate. There’s a physician in my area who will prescribe ketamine off-label to MDD and BP patients. A series of infusions are required; insurance generally won’t cover it. I am tempted to schedule a consult anyway. The vague, potential promise of “rapid relief” is just too enticing.

  23. Guy Chapman says:

    You mention EMDR. This is interesting to me. I do not know whether it works, whether it improves the efficacy of psychological treatment that would otherwise work anyway, or what, but I was offered it for PTSD and having first checked with a contact who happens to be an expert on the psychiatry of PTSD I gave it a go.

    My problem with PTSD was that the narrative surrounding the incident (a serious bicycling crash) was disjoint: playing the events in my head, I could not get past the feeling in the moments leading up tot he crash, of “I am going to die”, even though intellectually I know I didn’t. The EMDR therapy re-established the narrative thread so that I could observe the feeling I had had, but without re-experiencing it, because I could identify it as part of a narrative which culminated with me being strapped to a spinal board and eventually walking out of the hospital unaided later that day with nothing much worse than a broken rib.

    So, I can’t say what the role of the EMDR component was, but it does sound as if it was a proper, evidence-based use of it, in a relevant clinical context. Unfortunately, it’s also apparently being pushed by quacks and charlatans. My contact advised me that the evidence supports its use only where there is a clear and unambiguous diagnosis of PTSD. Importantly, I went in with a healthy skepticism and discussed this with the psychologist, who agreed that it is being oversold in some contexts. And “it worked for me” (those dread words), but it worked in the way the literature says it should and for the condition the literature lists as an indication.

    Bear in mind that this was after probably 8-10 sessions with David, a properly trained and eminently qualified psychologist with an honest-to-goodness PhD from a highly regarded department at a respected university and all, and the result was apparent (in reduced flashbacks, reduced nightmares, and reduced panic responses) after just two sessions. That is, normal psychotherapy has only a minor effect, EMDR had a dramatic and immediate effect, and subsequent psychotherapy again only a minor improvement (on a much better baseline).

    I hope I am not a dupe! I think we’d all like to know *how* this works, if it really does. As far as I can tell, my experience is fairly typical. So maybe it is just another psychologist’s mind game (David really did not like me using that term, but it’s what they are I reckon).

  24. Harriet Hall says:

    @Guy Chapman,

    There are studies showing that EMDR with psychotherapy is effective, but it’s far from clear that it offers any advantage over other treatments. My guess is that it works by distraction, or what they call “dual attention.” Our conscious attention is limited, and anything that reduces the amount of attention devoted to the emotional component of memories might be expected to facilitate psychotherapy. I would like to think there is a way to accomplish the same thing without deception and without overblown claims for a “gimmick.”

Comments are closed.