Text-only Preview

Journal of Psychopharmacology


A randomized, controlled pilot study of MDMA (3,4-Methylenedioxymethamphetamine)-assisted
psychotherapy for treatment of resistant, chronic Post-Traumatic Stress Disorder (PTSD)
Peter Oehen, Rafael Traber, Verena Widmer and Ulrich Schnyder
published online 31 October 2012
J Psychopharmacol
DOI: 10.1177/0269881112464827

The online version of this article can be found at:

Published by:


On behalf of:

British Association for Psychopharmacology

can be found at:
Journal of Psychopharmacology
Additional services and information for

Email Alerts:




- Oct 31, 2012
OnlineFirst Version of Record

What is This?
at University of Helsinki on November 16, 2012
Downloaded from

464827JOP0010.1177/0269881112464827Journal of PsychopharmacologyOehen et al.
Original Paper
A randomized, controlled pilot study of MDMA
assisted psychotherapy for treatment of
Journal of Psychopharmacology
0(0) 1 -13
resistant, chronic Post-Traumatic Stress
(c) The Author(s) 2012
Reprints and permission:
Disorder (PTSD)
DOI: 10.1177/0269881112464827
Peter Oehen1, Rafael Traber2, Verena Widmer1 and Ulrich Schnyder3
Psychiatrists and psychotherapists in the US (1970s to 1985) and Switzerland (1988-1993) used MDMA legally as a prescription drug, to enhance the
effectiveness of psychotherapy. Early reports suggest that it is useful in treating trauma-related disorders. Recently, the first completed pilot study
of MDMA-assisted psychotherapy for PTSD yielded encouraging results. Designed to test the safety and efficacy of MDMA-assisted psychotherapy in
patients with treatment-resistant PTSD; our randomized, double-blind, active-placebo controlled trial enrolled 12 patients for treatment with either
low-dose (25 mg, plus 12.5 mg supplemental dose) or full-dose MDMA (125 mg, plus 62.5 mg supplemental dose). MDMA was administered during three
experimental sessions, interspersed with weekly non-drug-based psychotherapy sessions. Outcome measures used were the Clinician-Administered PTSD
Scale (CAPS) and the Posttraumatic Diagnostic Scale (PDS). Patients were assessed at baseline, three weeks after the second and third MDMA session
(end of treatment), and at the 2-month and 1-year follow-ups.
We found that MDMA-assisted psychotherapy can be safely administered in a clinical setting. No drug-related serious adverse events occurred. We
did not see statistically significant reductions in CAPS scores (p = 0.066), although there was clinically and statistically significant self-reported
(PDS) improvement (p = 0.014). CAPS scores improved further at the 1-year follow-up. In addition, three MDMA sessions were more effective than
two (p = 0.016).
Methylenedioxymethamphetamine, MDMA, MDMA-assisted psychotherapy, psychotherapy, posttraumatic stress disorder, PTSD, entactogen
Post-traumatic stress disorder (PTSD) is a common problem in chronic illness, with high rates of psychiatric and medical comor-
everyday medical practice and a major and costly public health bidity (Jacobsen et al., 2001; McFarlane, 2010), as well as suici-
problem all over the world. Lifetime prevalence in the general dality (Panagioti et al., 2012). Serotonergic agents such as selective
population ranges from below 1% in the European countries serotonin re-uptake inhibitors (SSRI) and serotonin-norepineph-
(Hepp et al., 2005; Perkonigg et al., 2000), up to an average of rine reuptake inhibitors (SNRI) are often used to treat PTSD and
8% in countries such as the US (Breslau et al., 1991; Kessler comorbid disorders, or to treat patients unable to undergo psy-
et al., 1995), although more recent surveys in the Netherlands and
chotherapy. The only two FDA-approved drugs for this indica-
Switzerland now show rising rates of 7.4% in adults and 4.2% in tion, sertraline and paroxetine (Brady et al., 2000; Tucker et al.,
adolescents, respectively (De Vries and Olff, 2011; Landolt et al.,
2001), show only modest effects on PTSD symptoms. Recent
under review). In specific populations (e.g. soldiers returning literature reviews stress the importance of developing more effec-
from military service) prevalence can be much higher (Hoge tive medications and psychotherapeutic treatments for chronic
et al., 2004). Psychotherapy has been recognized to be the most PTSD (Foa et al., 2009; Stein et al., 2009).
effective form of treatment for PTSD (Van Etten et al., 1998).
MDMA (3,4-Methylenedioxymethamphetamine)-assisted psy-
First-line treatments are exposure-based therapies, such as: chotherapy is a novel approach to the treatment of PTSD that employs
Cognitive Behavior Therapy (CBT), Prolonged Exposure (PE),
Cognitive Processing Therapy (CPT) or Eye Movement
Desensitization and Reprocessing (EMDR) (Benedek et al., 2009;
1Private practice of Psychiatry and Psychotherapy, Biberist, Switzerland;
Cloitre, 2009; Foa et al., 2009). While demonstrating efficacy for 2Psychiatric Hospital, Marsens, Switzerland
some patients, studies of CBT show high drop-out rates (20%) 3Department of Psychiatry and Psychotherapy, University Hospital,
and limited effect on PTSD symptoms, with up to 58% of study Zurich, Switzerland
completers still meeting PTSD diagnosis after treatment, while Corresponding author:
only 32-66% reach a good level of end-state functioning (Foa, Peter Oehen, Private practice of Psychiatry and Psychotherapy,
2009; Schnyder, 2005). Despite a better understanding and grow-
Ulmenweg 24a, 4562 Biberist, Switzerland.
ing efficacy of the existing psychotherapies, PTSD often remains a Email: [email protected]
at University of Helsinki on November 16, 2012
Downloaded from

Journal of Psychopharmacology 0(0)
the psychoactive compound MDMA as a catalyst of PTSD-specific to traumatic memories should be kept in an "optimal arousal
psychotherapy itself. The drug MDMA is a substituted phenylethyl-
zone," avoiding the extremes of eliciting overwhelming anxiety
amine that was first synthesized in 1912 by the pharmaceutical com-
and other painful emotions that may lead to dissociation on the
pany Merck, but was rediscovered in the 1970s by the chemist A one hand and emotional numbing on the other (Ogden and Pain,
Shulgin, and it was later introduced to psychotherapy by the psycho-
2005). MDMA may widen this window, enhancing affect-related
therapist L Zeff (Benzenhoefer and Passie, 2006). Prior to the US tolerance and reducing numbing. The pronounced increases in
scheduling of MDMA as a drug of abuse in 1985, there were reports levels of the neurohormone oxytocin when patients are under the
suggesting it to be effective in psychotherapy (Greer and Tolbert, influence of MDMA (Wolff et al., 2006) are associated with the
1986; Metzner and Abramson, 2001). The first rigorously-controlled prosocial effects of MDMA (Bedi, 2009; Dumont et al., 2009).
clinical trials of MDMA-assisted psychotherapy in the treatment of The quality of this therapeutic alliance is recognized as being
chronic PTSD show promising results (Bouso et al., 2008; Mithoefer,
crucial for the recovery from PTSD (Charuvastra and Cloitre,
2011). The benefits of MDMA-assisted psychotherapy appear to be 2008); while the extensive release of oxytocin under MDMA
long-lasting (Mithoefer et al., in press).
therapy is postulated to be a prominent factor in improving
The current neurocircuitry model of PTSD postulates that the therapeutic alliance that is regularly observed in clinical-
there are exaggerated and uncontrolled responses of the amyg-
therapeutic settings, in patients under the influence of MDMA
dala to trauma-specific cues, as well as a deficient top-down (Johansen and Krebs, 2009). The main postulated psychological
inhibition of the amygdala by the ventromedial prefrontal cor-
effects that are relevant to the context of MDMA-assisted psy-
tex (vmPFC), orbitofrontal cortex and the hippocampus chotherapy are partially based on clinical impressions and also
(Frewen, 2006; Rauch et al., 2006). MDMA increases activity on clinical data (Johanson and Krebs, 2009; Passie and Durst,
in the vmPFC and decreases the activity of the left amygdala 2009; Vollenweider et al.,1998), as is shown in Table 1.
(Gamma et al., 2000), possibly reversing some of the above-
This study was intended to serve as a proof of concept and to
mentioned abnormalities associated with PTSD. MDMA leads secondarily confirm the initial findings of the Mithoefer et al.
to a transporter-mediated release of serotonin and the activa-
(2011) study, using a different therapist team. We examined safety
tion of the 5HT receptor, and to a lesser extent to the release of and efficacy in an outpatient setting, which included overnight
dopamine and norepinephrine. Many of the positive subjective stays after each MDMA session in the clinic, for safety reasons, in
effects can be attributed to the release of serotonin (Farre et al.,
a small sample of 12 patients with chronic, treatment-resistant
2007; Liechti et al., 2001), and it was recently shown for nor-
PTSD and provided a 1-year follow-up. A methodological chal-
epinephrine, as well (Hysek et al., 2011). A main characteristic lenge was the maintenance of the double-blind, when using a pro-
of the MDMA-induced state is a positively-toned cognitive-
foundly psychoactive substance like MDMA, since MDMA's
emotional state, with reduced fear, possibly facilitating the effects can be easily discerned by both subjects and investigators.
processing of traumatic material and better encoding positive In the Mithoefer et al. study (2011), there were difficulties in
emotional experiences. It is theorized that therapeutic exposure
maintaining the study blind using an inactive placebo control, so
Table 1. Psychological effects of MDMA in the context of psychotherapy.
MDMA-induced state
Psychotherapeutic implication
Mild euphoria
Positive and fearless emotional state of well-being
Anxiety and fear
Emotional avoidance
Enhanced perception of and intensified
Tolerance and processing of difficult emotions
("window of tolerance")
Affect tolerance

Cognition/ Memory
More imaginative and associative
Recall of relevant traumatic memories
Prolonged spontaneous exposure to traumatic
Recall and tolerance of traumatic memories
Cognitive restructuring
"simulation of alternative behavior"
Social fears and defensiveness
Improvement of therapeutic alliance
Interpersonal Behavior
Social approach behavior with empathy,
Rebuilding of trusting relationships
openness, trust, feelings of being connected
Defensiveness and isolation
to others
Cuddling and need for touch

Consolidation of self
Release of muscular tension
Release of tension and reduction of somatic

Positive body image
MDMA: 3,4-methylenedioxymethamphetamine
at University of Helsinki on November 16, 2012
Downloaded from

Oehen et al. 3
it is possible that this difficulty affects study results. This study, included significant medical conditions, except for hypothyroidism
therefore, attempted to address the question of whether the use of under hormonal replacement. Exclusionary psychiatric conditions
25mg of MDMA as an "active placebo" could help optimize were: a history of psychotic illness, bipolar disorder type I, bor-
blinding. We also hypothesized that three MDMA sessions were derline personality disorder, dissociative identity disorder, and
more effective than only two, and that reductions in PTSD symp-
substance abuse or dependence within 60 days of enrollment. We
toms would remain stable at the 1-year follow-up.
did allow comorbid anxiety disorders, depression, and eating dis-
orders without active purging. We excluded subjects who had
taken MDMA on more than five occasions or less than 6 months
prior to enrollment.
Recruitment and screening procedure
As far as previous drug use, one subject had previously
used "ecstasy" on three occasions, one had consumed magic
We recruited subjects for the study by calling for referrals from mushrooms (psilocybin) several times, while the other sub-
psychiatric hospitals, trauma counseling centers, psychia-
jects were completely naive to psychedelic drugs. Two sub-
trists and psychotherapists in the German-speaking part of jects (one female, one male) discontinued treatment after the
Switzerland. First, prospective participants were screened by a first experimental MDMA session. Eleven of the 12 subjects
scripted telephone interview, to check for inclusion and exclu-
who completed the study also participated in the 12-month
sion criteria. Those who met the study criteria had an informa-
follow-up. One female subject could not complete the
tional meeting with the investigator, which included the 12-month follow-up, because she died 6 months after finish-
administration of the Clinician-Administered PTSD Scale ing the MDMA-assisted treatment, from a brain metastasis
(CAPS) to provide a PTSD diagnosis. Written informed consent
arising from a relapse of breast cancer; when chosen for inclu-
was then obtained from subjects by the investigators. Medical sion, this subject had been in breast cancer remission for over
evaluation included: a medical history, standard physical exam-
10 years and had not been symptomatic at screening.
ination, an electrocardiogram (ECG), metabolic profile, meas-
Index traumata within the study group included physical and
urement of thyroid hormones, serum electrolytes, human sexual abuse during childhood in six subjects, sexual assault in
immunodeficiency virus (HIV) test, urinary drug test and preg-
one, medical treatment in one, motor vehicle accident in two and
nancy test (when appropriate). The subjects aged older than 40 life-threatening illness in two subjects. The mean duration of
years, with a positive family history of coronary heart disease PTSD symptoms at enrollment was 18.3 years (SD 12). The
and/or presenting risk factors, underwent a stress ECG. mean duration of previous psychotherapeutic treatments was 85.8
Psychiatric evaluation and confirmation of the PTSD diagnosis months (SD 71.4).
were conducted by an independent rater, using CAPS and
Subjects were allowed to continue any ongoing psychotherapy
Structured Clinical Interview for Diagnostic and Statistical with their outside/referring therapists, but were not allowed to
Manual of Mental Disorders (DSM) (SCID) I and II. Enrollment
increase the frequency of the ongoing treatments, nor commence
began in September 2006 and ended in October 2009. The any new therapy, until after the administration of outcome meas-
12-month follow-up was completed in January 2011. This study
ures at 2 months after the MDMA session #3.
was approved by the ethics committee of the cantons (federal
states) of Solothurn and Aargau, Switzerland and it was con-
ducted according to the regulatory guidance for protection of Description of study design
human subjects, and relevant federal regulations and interna-
In "Stage 1," eight subjects were randomized in a double-
tional standards.
blind manner to the full dose and four to the "active placebo"
condition, with their three doses of MDMA administered in
three all-day-long MDMA-assisted psychotherapy sessions.
A full dose consisted of 125 mg followed 2.5 hours later by
We enrolled 12 subjects (10 female, two male; mean age = 41.4 62.5 mg MDMA; while the "active placebo" dose consisted
years, SD 11.2 years) meeting all inclusion and exclusion criteria, of 25 mg, followed 2.5 hours later by 12.5 mg MDMA. The
whom completed the study (Table 2). Two additional subjects dis-
125 mg dose of MDMA was chosen on the basis of case
continued treatment after the first experimental MDMA session reports of MDMA-assisted psychotherapy (Greer and Tolbert,
(Figure 1). All subjects who were enrolled, met the DSM-IV-text 1986; Widmer, 1998), as well as on preliminary data obtained
revision (TR) criteria for PTSD with treatment-resistant symp-
from the Mithoefer (2011) pilot study. The dosages chosen for
toms, as was indicated by a CAPS score of 50 and having previ-
the low dose condition were selected on the basis of their
ously undergone at least 6 months of psychotherapy and 3 months
ability to produce minimal, but detectable subjective effects
of treatment with an SSRI. Seven of 12 subjects had experienced (Grob et al., unpublished; Harris et al., 2002), thus serving as
one or more evidence-based therapies: three subjects had CBT, an "active placebo." The cumulative dose of 37.5 mg MDMA
one exposure based therapy was not specified, one had EMDR, was not expected to produce a significant reduction in anxiety
three had unspecified anxiety management and six subjects had nor a significant increase in access to emotionally-upsetting
non-evidence based therapies, such as insight-oriented therapies. material, although this low dosage might produce slight alter-
Many of the subjects had undergone multiple therapies, so it was ations in perception, and increased relaxation or tension
not possible any more for them to exactly identify in all cases, the (Harris et al., 2002). The study purposefully allocated a
specific method that had been applied. Subjects were required to greater number of participants to the full-dose condition
taper all psychotropic medication, before entering the study. (2:1), to better assess the safety of the full dose and to enhance
Gabapentin was allowed for pain control. Exclusion criteria recruitment efforts.
at University of Helsinki on November 16, 2012
Downloaded from

Journal of Psychopharmacology 0(0)
Figure 1. Flow Diagram.
Outcome measures
version of the CAPS was used (Schnyder et al., 2002), serving as both
a screening and main outcome measure.
Outcome measures included two measures of PTSD symptoms.The
The Posttraumatic Diagnostic Scale (PDS) (Foa et al., 1993,
Clinician-Administered PTSD Scale (CAPS) is a DSM-IV based, 1997) is a validated self-reporting measure to assess the presence
structured clinical interview that is designed to quantify PTSD symp-
of PTSD symptoms, as is described in the DSM-IV serving as an
toms. It was determined to have excellent psychometric properties of additional outcome measure. An unvalidated, yet widely-used
reliability and validity (Weathers et al., 2001). A validated German German version (Ehlers et al., 1996) was used in this study.
at University of Helsinki on November 16, 2012
Downloaded from

Oehen et al. 5
Table 2. Study participant characteristics.
Full-dose group
Placebo group

n = 8
n = 4
n = 12
7 (87%)
3 (75%)
10 (83%)

1 (12%)
1 (25%)
2 (16%)
Mean age (SD)
Range 23-67 yrs
42.1 (12.8)
40.0 (6.2)
41.4 (11.2)
Country of origin
Study completers
CH: 7, F: 1
CH: 4
CH: 11, F: 1

TR: 1
ZA: 1

Marital status
3 (37%)
2 (50%)
5 (41%)

Married/living with partner
2 (25%)
2 (50%)
5 (41%)

3 (37%)
0 (0%)
4 (33%)
Work status
On disability
4 (50%)
1 (25%)
5 (42%)

Fit for limited employment
2 (25%)
1 (25%)
3 (25%)

Working full-time
1 (13%)
2 (50%)
3 (25%)

1 (13%)
0 (0%)
1 (8%)
History of abuse/dependency
1 (13%)
0 (0%)
1 (8%)

1 (13%)
1 (25%)
2 (17%)
Prior drug use
MDMA (# subjects)
1 (3 occasions)

Psilocybin (# subjects)
Mean # years duration of
Range 3-40 y
16.4 (10.9)
22.3 (12.1)
18.3 (12.0)
Mean # months of prior

Range 22-240 m
39.9 (73.3)
123 (60.6)
85.8 (71.4)
psychotherapy (SD)
Comorbid disorder

Unipolar depression
7 (88%)
3 (75%)
10 (83%)

Panic disorder
0 (0%)
1 (25%)
1 (8%)

Eating disorder
1 (13%)
0 (0%)
1 (8%)

Seasonal affective disorder
1 (13%)
1 (25%)
2 (17%)

1 (13%)
0 (0%)
1 (8%)
Index trauma
Childhood sexual abuse
4 (50%)
2 (50%)
6 (50%)

Sexual assault
1 (13%)
0 (0%)
1 (18.%)

1 (13%)
1 (25%)
2 (17%)

Medical treatment
1 (13%)
0 (0%)
1 (28%)

Life-threatening illness
1 (13%)
1 (25%)
2 (17%)
Medication for PTSD at
4 (50%)
2 (50%)
2 (50%)
CH: Switzerland; F: France; MDMA: 3,4-Methylenedioxymethamphetamine ;TR: Turkey; ZA: South Africa
The CAPS and SCID I substance abuse module were adminis-
"Stage 2." All subjects in the "active placebo" condition in "Stage
tered at baseline (T0), 3-weeks after MDMA-session #2 (T1); 1" chose to proceed to "Stage 2." Follow-up assessments consist-
3-weeks after MDMA-session #3 (T2; end of treatment); and two ing of the CAPS and PDS were completed two (T3), six (T4) and
(T3), six (T4) and 12 (T5) months after the MDMA-session #3 12 (T5) months after the final MDMA-session #3.
(follow-up). The PDS was administered one day after each MDMA
After a preliminary analysis of data showed an insufficient
session; 3-weeks after the MDMA-session #3 (T2; end of treat-
clinical response to the experimental treatment in several full-
ment); and two, six, and 12 months after MDMA-session #3 (T3, dose subjects, an amendment to the protocol was obtained, allow-
T4, T5; long-term follow-up (LTFU)). All outcome measures were ing for two additional sessions of MDMA-assisted psychotherapy
administered by a blinded, independent rater. Subjects were tested for any subjects deemed to show insufficient response, which was
for drugs of abuse before MDMA sessions, plus 1-time at random, referred to as "Stage 3" and employed a dose of 150 mg MDMA
during Stage 1 and Stage 2, and at each follow-up testing. and a supplemental dose of 75mg MDMA, unless contraindicated
Pregnancy tests were performed in women of childbearing poten-
for safety reasons. A response was considered clinically insuffi-
tial, before each MDMA session, as a safety measure. The blind cient on the basis of:
was broken following assessment by the independent rater, after
the end of Stage 1 treatment. Subjects assigned to the "active pla-
- the investigator's and patients' subjective impression of a
cebo" condition were offered an open-label continuation of the
lack of improvement
study with the fully active dose of MDMA ("Stage 2"), with identi-
- CAPS score changes (baseline to 2 months after the third
cal psychotherapy and assessment as in "Stage 1." CAPS scores
experimental session 15 points (Schnurr, 2007; Weathers,
from the 3-weeks post-MDMA #3 testing served as a baseline for
at University of Helsinki on November 16, 2012
Downloaded from

Journal of Psychopharmacology 0(0)
- CAPS item #25 3 and overall CAPS score still 50 sessions. The therapists' attitude was supportive, validating the
points at the outcome measurement 2-months after the third
MDMA experience and facilitating understanding and emotional
MDMA-session served as additional guidelines for the clearing. Following each MDMA-assisted session, the subjects
assessment of clinically insufficient response).
were contacted via telephone by one of the therapists on a daily
basis for one week, in order to assess the subject's psychological
All three above conditions had to be fulfilled.
well-being and monitor any drug after-effects. Subjects each
received a total of 12 non-drug psychotherapy sessions. Additional
sessions in case of excessive distress were limited to two after
each MDMA session.
MDMA was obtained from a supply originally synthesized by
Lipomed AG, Switzerland. The investigational product (in quan-
Further assessment and safety measures
tities of 125, 62.5, 25 and 12.5 mg) was prepared in gelatin cap-
sules of identical appearance and weight by the Bichsel Subjects' blood pressure (BP) and heart rate (HR) were measured
Laboratory in Interlaken, Switzerland. Quality control and rand-
at both15 and 5 minutes before ingestion of the MDMA, and after-
omization was performed by R Brenneisen, Department of wards every half-hour for a total of 4 hours and then every hour
Clinical Research, Phytopharmacology, Bioanalytics and until the termination of the session. Body temperature was meas-
Pharmacokinetics, University of Bern, Switzerland.
ured 15 minutes before MDMA administration and hourly, until
termination of the session. The degree of psychological distress
was monitored repeatedly during the course of each MDMA ses-
sion, using a 1-item visual analog scale, the Subjective Units of
Distress. The participant's beliefs concerning their condition were
The treatment is described online in the manual for MDMA-
collected during the non-drug psychotherapy session given the
assisted psychotherapy in patients with PTSD (Mithoefer, 2011). day after each MDMA session. The therapists collected any spon-
Two preparatory sessions, aimed at establishing a therapeutic alli-
taneously-reported reactions over a 7-day period, starting on the
ance and preparing subjects for the MDMA experience, preceded day of each experimental session.
the first MDMA session of the study. The MDMA sessions took
place in the group psychotherapy room at the first author's clinic.
Subjects arrived at nine a.m. After testing for drugs of abuse and Statistical analysis
testing of females for pregnancy, the session goals and intentions
were recapitulated. The MDMA was ingested at 10 a.m. Subjects CAPS and PDS scores were analyzed by nonparametric analysis of
were instructed to remain reclining on the mattress, to focus their variance (ANOVA), using an F1-LD-F1 model (Brunner and
attention inward, keep their eyes closed as much as possible and Langer, 1999; Brunner et al., 2002) with the experimental interven-
to allow the inner process to unfold. The therapeutic tools used to tion condition (full dose MDMA versus "active placebo" MDMA)
guide the subjects consisted of:
serving as a between-group factor and the time of measurement
serving as a within-subjects factor. Given an insufficient number of
1. A program of music which was designed to support the sub-
participants in "Stage 2" for formal analysis, their scores were
ject's experience by aiding relaxation and/or evoking and compared across the two stages to see whether "Stage 2" scores
supporting deep emotions and the emergence of uncon-
were reduced, as compared to "Stage 1" scores. The Wilcoxon
scious material (Bonny and Savary, 1990; Spitzer, 2002).
Signed-Rank-Test for paired data was used to analyze whether a
2. MDMA-assisted psychotherapy is primarily focused on third MDMA session improved CAPS scores compared to only
experiencing and is only to a lesser extent a verbal method
two MDMA sessions. Group comparisons of vital signs pre- to
during the MDMA sessions themselves. Discussions post-session (excluding data from the high-dose group, due to
between therapists and participant take place only when insufficient sample size) were performed by first averaging the val-
needed. The therapeutic approach is generally non-direc-
ues for each subject over the three sessions to obtain an "average"
tive, following and encouraging the MDMA-induced day and then calculating a nonparametric 95% confidence interval
psychological process.
(CI) covering the true median of the differences pre- to post-ses-
3. Focused body work was defined as bodily contact that sion. To compare the magnitude of the difference between the
employs nurturing touch (e.g. hand-holding) and touch maximally-observed value and the baseline value between treat-
aimed at intensifying and thereby releasing body tension and
ment groups, a similar approach as above was chosen: To show
pain, by giving resistance for the subject to push against. It is
that the values of the increase are higher on average in the full-dose
always performed with explicit consent from the subject and
group than in the placebo group, a lower confidence bound B for
respecting individual boundaries and vulnerabilities.
the difference of increase, such that the true value of increase (full
dose) - increase ("active placebo") is at least as big as B with a
One male and one female therapist were were present during confidence of 95%, was computed. Given the small sample size,
the entire session. MDMA-assisted sessions lasted approximately no adjustments for covariates were made and the study had only
8 hours, after which the subjects were offered a light meal, and sufficient power to detect large effects; therefore, there was no
then a previously-designated support person (e.g. spouse) arrived adjustment for multiple testing: unadjusted exact p-values and CIs
to stay with them overnight at the clinic. A non-drug psychother-
were reported instead. Results were considered significant when
apy session took place the morning after each MDMA experience,
p 0.05. Trends were also reported when p 0.1. The F1-LD-F1
followed by two sessions that were one week apart, aimed at ensur-
models were computed with the SAS 9.1 program; all other analy-
ing the integration of the experiences from the MDMA-assisted ses were performed with the R 2.7.1. statistics program.
at University of Helsinki on November 16, 2012
Downloaded from

Oehen et al. 7
CAPS mean total scores by group for me TO-T2 (SD)
PDS mean scores by group for me TO-T2 (SD)
TO: Baseline < 4 weeks before MDMA and a er
TO: Baseline < 4 weeks before MDMA and a er
discon nua on of psychtropic medica on
discon nua on of psychtropic medica on
Ac ve Placebo: 63.4 (7.9) Full Dose: 66.4 (13.6)
Ac ve Placebo: 23.5 (1.9) Full Dose: 30 (6.3)
T1: 3 weeks post MDMA-session 2
Ac ve Placebo: 60.0 (6.8) Full Dose: 63.0 (17.8)
T2: 3 weeks post MDMA-session 3
T2: 3 weeks post MDMA-session 3
(end of treatment)
(end of treatment)
Ac ve Placebo: 30.8 (6.2) Full Dose: 21.4 (11.9)
Ac ve Placebo: 66.5 (7.6) Full Dose: 50.8 (19.7)
CAPS Change scores (SD):
Changes scores T0-T2 (SD):
T0-T1: Ac ve Placebo: -3.3 (9.9) Full dose: -3.4 (12.0)
Ac ve Placebo: 7.3 (6.2) Full dose: -8.6 (13.0)
T1-T2: Ac ve Placebo: 6.5 (10.3) Full dose: -12.2 (8.1)
T0-T2: Ac ve Placebo: -3.2 (15.3) Full dose: -15.6 (18.1)
Figure 2. CAPS and PDS scores by group for time T0-T2.
the full-dose group, as compared to an increase in the "active
placebo" group. There was a significant interaction effect of
group and time (p = 0.014).
A Wilcoxon Signed Rank test for paired data was performed to
Figure 2 shows the course of CAPS and PDS scores over time test whether three MDMA sessions were more effective than only
in the two groups. Interestingly, the average CAPS scores in two sessions. There was a significant difference in CAPS scores
the "active placebo" group increased slightly from T1 to T2. (p = 0.016, exact p-value to account for ties) between the two time
The three interaction relative treatment effects (RTE) T0-T2 points T1 and T2.
for total CAPS scores in the full dose group showed a distinct
The median prior psychotherapy treatment times of the "active
decrease in CAPS scores with time, as compared to the active placebo" and the full-dose group were 123 and 39.9 months. A
placebo group in the ANOVA, but narrowly missed statistical comparison of the two distributions, using the two-sample
significance (p = 0.066). On average, CAPS scores decreased Wilcoxon rank sum test, yielded a two-tailed p value of 0.154.
15.6 points (23.5%) in the full-dose subjects. There was a sig-
nificant simple effect of time in the full dose group (p = 0.002),
meaning that the time effect was significant only in the full-
dose group. In contrast, the simple time effect for the active
placebo group was not significant (p = 0.475). For the other There were no serious drug-related adverse events and no medical
two models, T0 vs. T1 and T1 vs. T2, group and time effects intervention was required during or following the MDMA
and interactions were not significant. PDS scores decreased in sessions.
at University of Helsinki on November 16, 2012
Downloaded from

Journal of Psychopharmacology 0(0)
Table 3. Physiologic data.
MDMA Group
Full Dose (excl. high dosea)
"Active" Placebo (low dose)

Mean/ median
Mean/ median
Systolic BP


Diastolic BP






aSample size too small for statistical analysis; all values for BP, HR and T were within ranges of the full-dose group.
BP: blood pressure; HR: heart rate or pulse; MDMA: 3,4-Methylenedioxymethamphetamine; T: temperature
Rescue medication
reported in the active placebo group. Dizziness, headache and
impaired gait/balance were also frequently reported in both groups.
Zolpidem for insomnia was offered for the first nights after MDMA
Most reactions resolved when the drug effects diminished. Loss of
sessions, but was administered on only one occasion. Most subjects appetite, difficulty concentrating, anxiety and headache persisted
refused sleep medication, frequently commenting that lying awake beyond this window to 24 hours, but were still self-limiting.
was not experienced as being distressing, but an opportunity to
reflect on the still ongoing inner process. Lorazepam for anxiety/
distress related to the processing of the traumatic memories was Physiologic data
administered in six out of nine subjects, after 10 out of 56 full-dose
or 150 mg MDMA sessions, typically during the week after MDMA
As seen in Table 4, for both groups, the temperature values tended
sessions. Five of these six subjects were on antidepressants and/or to be significantly higher pre- to post-session, the increase being
benzodiazepines at enrollment. In all cases, single doses of 1-2mg within the range between 0.97 and 0.46 degrees Celsius. In the
lorazepam reduced the anxiety or distress adequately. Only one sub-
full-dose group, systolic BP and HR did not change significantly
ject, with no psychotropic medication at enrollment, required loraz-
(albeit just narrowly, which may be due to underpowering). The
epam on one occasion. In the "active placebo" group, lorazepam comparison of the difference between the maximally-observed
was administered to two of five subjects after three low-dose and the baseline value between conditions showed that all lower-
MDMA sessions. Both had been treated with antidepressants and/or
confidence bounds B were negative, meaning the increase in any
benzodiazepines at enrollment. The other three active placebo sub-
of the physiological parameters was not significantly higher in the
jects did not need any medication, nor had they had any psycho-
full-dose than in the placebo group.
tropic medication at enrollment. Except for the subject who was
subsequently diagnosed with a prefrontal brain metastasis and who
experienced a panic attack, the anxiety that required medication Additional psychotherapy sessions
was related to the PTSD. Acetaminophen or mefenamic acid (in Additional integrative psychotherapy sessions were conducted as
two subjects with a history of headache refractory to acetami-
per protocol in situations of excessive distress or other issues, fol-
nophen) were administered short-term for headache, following the lowing MDMA sessions. Eight out of 13 subjects who received full-
MDMA sessions.
dose either in the initial randomization or in the "Stage 2" crossover
group, required a total of 21 additional sessions, with no more than
Spontaneously-reported reactions
four additional sessions per subject and stage (mean 1.6 per sub-
ject). In the "active placebo" group (n = 5), four additional sessions
See Table 3. The most commonly reported reactions on the day of were provided to the above-mentioned two "active placebo" sub-
jects exhibiting excessive distress (mean 0.8 per subject). One addi-
the experimental session were moderate insomnia (125 mg: 43%; tional session was conducted in "Stage 3" (mean 0.3 per subject).
150 mg: 50%), loss of appetite and restlessness in subjects receiving
125 mg MDMA, and headache, moderate insomnia (31%) and loss
of appetite in subjects receiving 25 mg MDMA. Insomnia and loss Clinical response and LTFU
of appetite were the most commonly reported reactions in both con-
ditions. Restlessness, tight jaw, thirst and feeling cold were com-
Clinical response, as defined above, was observed in four out of
monly reported reactions in the full-dose group that were minimally eight subjects in the full-dose group, with all of them still
at University of Helsinki on November 16, 2012
Downloaded from

Oehen et al. 9
Table 4. Spontaneously-reported reactions.
Day of MDMA
Day of Placebo
Day of MDMA
Within 7 days
Within 7 days
after MDMA
after low dose

125 mg
25 mg
150 mg
125/150 mg
25 mg

Sessions: 37
Sessions: 13
Sessions: 6
Sessions: 43
Sessions: 13

n (%) (Mean
n (%) (Mean
n (%) (Mean
n (%) (Mean
n (%) (Mean
10 (27%) (1.6)
2 (15%) (1.4)
1 (16%) (1.0)
11 (26%) (1.0)
2 (15%) (1.0)
Decreased concentration
6 (16%) (1.1)
10 (23%) (1.5)
8 (22%) (1.0)
4 (31%) (1.0)
3 (50%) (2.3)
8 (18%) (1.6)
3 (23%) (1.3)
2 (5%) (1.0)
2 (5%) (1.0)
1 (8%) (1.0)
Dry mouth
7 (19%) (1.1)
2 (33%) (1.5)
5 (12%) (1.0)
13 (35%) (1.5)
2 (15%) (1.0)
1 (16%) (1.0)
24 (56%) (1.6)
5 (38%) (1.4)
11 (30%) (1.6)
5 (38%) (1.8)
2 (33%) (1.5)
10 (23%) (1.9)
4 (31%) (1.5)
Heavy legs
1 (3%) (1.0)
1 (16%) (1.0)
1 (8%) (1.0)
Impaired gait/balance
12 (32%) (1.0)
3 (23%) (1.0)
4 (66%) (1.0)
3 (7%) (1.4)
9 (21%) (1.3)
1 (8%) (1.0)
Increased private
2 (5%) (1.5)
9 (21%) (1.4)
3 (23%) (1.1)
16 (43%) (2.1)
4 (31%) (1.8)
3 (50%) (2.3)
20 (47%) (1.9)
6 (46%) (1.6)
Jaw clenching
14 (38%) (1.4)
1 (8%) (1.0)
4 (66%) (2.3)
7 (16%) (1.2)
Lack of appetite
15 (41%) (1.9)
4 (31%) (2.0)
2 (33%) (2.0)
7 (16%) (1.5)
5 (38%) (1.5)
Low mood
4 (11%) (1.3)
1 (8%) (2.0)
20 (47%) (1.4)
6 (46%) (1.4)
6 (16%) (1.8)
2 (15%) (1.0)
2 (33%) (1.0)
5 (12%) (1.0)
2 (15%) (1.2)
Need for more sleep
1 (3%) (2.0)
6 (14%) (1.1)
3 (23%) (1.2)
3 (8%) (1.0)
1 (16%) (1.0)
1 (2%) (1.0)
2 (5%) (1.0)
1 (16%) (1.0)
6 (16%) (1.5)
2 (33%) (1.0)
1 (2%) (1.0)
15 (41%) (1.2)
2 (33%) (1.5)
6 (14%) (1.4)
Feeling cold
11 (30%) (1.1)
1 (8%) (1.0)
6 (14%) (1.2)
1 (8%) (1.0)
13 (35%) (1.3)
2 (33%) (1.5)
1 (2%) (1.2)
3 (8%) (1.8)
1 (16%) (1.0)
5 (12%) (1.1)
n: Number of spontaneous reports;(%); n in percentage of sessions;
Severity: 1 = mild, 2 = moderate, 3 = severe;
MDMA: 3,4-Methylenedioxymethamphetamine
fulfilling PTSD criteria, but with a reduction in severity from decrease (52%) in the crossover group, with nine subjects show-
severe to mild (CAPS score 20-39) (n = 3) or moderate (CAPS ing a significant clinical improvement. During this time, the
score 40-59) (n = 1) PTSD.
majority of subjects continued with their previous or another psy-
Three full-dosage subjects met criteria for being non-responders
chotherapy or medication. Also at LTFU, five of 12 subjects no
and they were enrolled in "Stage 3," receiving either a full or higher
longer met the diagnostic criteria for PTSD, two had switched to
dose of MDMA (two full-dose sessions, two high-dose sessions having mild PTSD, and four had moderate PTSD, while one had
and two high-dose sessions followed by a lower supplemental died of a cause not related to the study. One of four subjects on
dose). The dosages were chosen on the basis of clinical judgment. disability and three who were fit for limited employment at base-
The additional sessions did not lead to any further improvements line had been able to return to work full-time by the 1-year
in CAPS scores (mean CAPS score change of 0.3 points). As a follow-up.
result, no further subjects were enrolled in "Stage 3."
In the "active placebo" group all four subjects failed to respond
to the treatment, with two subjects showing higher CAPS scores Blinding
and a slight clinical deterioration. In the "Stage 2" crossover The investigator's guesses on the 14 subjects' condition assign-
group, all four subjects responded to the treatment: two of four ments were correct in eight of the full-dose subjects (including
subjects no longer fulfilled PTSD criteria and two had improved, one drop-out) and uncertain in one full-dose subject. They were
but still had moderate PTSD. At the one-year follow-up, CAPS also correct in two of the active placebo subjects, whereas their
scores had decreased by a mean of 24 points (35%) compared to guesses were incorrect in one and they were uncertain in two of
baseline in the full-dose group, while there was a 35-point the active placebo cases (including one drop-out). A total of 13
at University of Helsinki on November 16, 2012
Downloaded from