RETRACTED ARTICLE: MDMA-assisted psychotherapy for treatment of PTSD: study design and rationale for phase 3 trials based on pooled analysis of six phase 2 randomized controlled trials

Background Posttraumatic stress disorder is a prevalent mental health condition with substantial impact on daily functioning that lacks sufficient treatment options. Here we evaluate six phase 2 trials in a pooled analysis to determine the study design for phase 3 trials of MDMA-assisted psychotherapy for PTSD. Methods Six randomized, double-blind, controlled clinical trials at five study sites were conducted from April 2004 to February 2017. Active doses of MDMA (75–125 mg, n = 72) or placebo/control doses (0–40 mg, n = 31) were administered to individuals with PTSD during manualized psychotherapy sessions in two or three 8-h sessions spaced a month apart. Three non-drug 90-min therapy sessions preceded the first MDMA exposure, and three to four followed each experimental session. Results After two blinded experimental sessions, the active group had significantly greater reductions in CAPS-IV total scores from baseline than the control group [MMRM estimated mean difference (SE) between groups − 22.0 (5.17), P < 0.001]. The between-group Cohen’s d effect size was 0.8, indicating a large treatment effect. After two experimental sessions, more participants in the active group (54.2%) did not meet CAPS-IV PTSD diagnostic criteria than the control group (22.6%). Depression symptom improvement on the BDI-II was greatest for the active group compared to the control group, although only trended towards significant group differences [MMRM, estimated mean difference (SE) between groups − 6.0 (3.03), P = 0.053]. All doses of MDMA were well tolerated, with some expected reactions occurring at greater frequency for the active MDMA group during experimental sessions and the 7 days following. Conclusions MDMA-assisted psychotherapy was efficacious and well tolerated in a large sample of adults with PTSD. These studies supported expansion into phase 3 trials and led to FDA granting Breakthrough Therapy designation for this promising treatment. Trial registration ClinicalTrials.gov Identifier: NCT00090064, NCT00353938, NCT01958593, NCT01211405, NCT01689740, NCT01793610. Electronic supplementary material The online version of this article (10.1007/s00213-019-05249-5) contains supplementary material, which is available to authorized users.


Introduction
Posttraumatic stress disorder (PTSD) is a serious debilitating disorder with lifetime prevalence estimated at nearly 4% globally and over 8% in the USA (Kilpatrick et al. 2013;Koenen et al. 2017). Symptoms of PTSD include intrusive thoughts and memories, negative effects on cognition and mood, hyperarousal and reactivity, and avoidance that do not remit for at least 1 month subsequent to exposure to a traumatic event (Koenen et al. 2017). Individuals with PTSD may experience a substantial reduction in quality of life and relationships, and the disability resulting from PTSD can have further negative consequences such as obesity (Scott et al. 2008), hypertension (Kibler et al. 2009), comorbid mental health conditions, and suicidality (Dorrington et al. 2014;Tarrier and Gregg 2004). In addition to these profound costs to individuals with PTSD, the disorder also exerts a substantial economic toll through lost productivity and treatment costs (Marshall et al. 2000).
Widely used treatments for PTSD include psychotherapies and medications. A recent review identified trauma-focused psychotherapies as first-line treatments for PTSD (Lee et al. 2016); however, while a substantial proportion of individuals with PTSD respond to psychotherapies [e.g., cognitive processing therapy (Monson et al. 2006;Resick et al. 2008) and prolonged exposure therapy (Foa et al. 2007)], these therapies may be difficult to access and are ineffective for many (Koenen et al. 2017;Steenkamp et al. 2015). A variety of medications have also been used to address PTSD symptoms, but only two drugs-sertraline and paroxetine-are approved by the FDA for PTSD. Extant pharmacotherapies, however, are ineffective for many individuals with PTSD, with an estimated 40-60% of patients not responding adequately (Bradley et al. 2005;Brady et al. 2000;Steenkamp et al. 2015). They may have problematic side effects and generally require long-term use to maintain effectiveness (Lee et al. 2016). In sum, the sizable proportion of cases of PTSD are persistent (Koenen et al. 2017) and the shortcomings of currently available treatments make the development of novel PTSD treatments a research priority.
A promising approach to the treatment of PTSD is the combination of psychotherapy with pharmacotherapy using 3,4-methylenedioxymethamphetamine (MDMA). Interest in the therapeutic potential of MDMA for trauma-related psychopathology developed in the context of the broader potential for MDMA to catalyze psychotherapeutic processes by facilitating communication and connection between therapists and patients (Nichols 1986). MDMA was first synthesized in 1912 by Merck, but it was not until the early 1970s that MDMA was first used in combination with psychotherapy. Case reports from that period described therapeutic benefits, although no clinical trials were conducted at that time. Recreational use of BEcstasy,^tablets purported to contain MDMA, became popular in the 1980s, leading to its classification as a Schedule 1 controlled substance in 1985. The scheduling of MDMA made its use in therapy illegal and created obstacles to clinical research. A non-profit organization, the Multidisciplinary Association for Psychedelic Studies (MAPS), filed a Drug Master File (DMF) application in 1986, followed by an Investigational New Drug (IND) application in 2001, embarking on the FDA drug development process to study the safety and efficacy of MDMA as an adjunct to psychotherapy for PTSD (Greer and Tolbert 1986;Grof 2001;Mithoefer 2011Mithoefer , 2017Mithoefer et al. 2018).
After nonclinical toxicity studies and an investigatorinitiated phase 1 study of MDMA were completed (Frith et al. 1987;Grob et al. 1996Grob et al. , 1998, six phase 2 randomized trials of MDMA-assisted psychotherapy for treatment of PTSD were conducted from 2004 to 2017. Active doses of MDMA (75-125 mg) or control doses of inactive placebo or low-dose MDMA (25-40 mg) were combined with manualized inner-directed psychotherapy (Mithoefer 2017) in which participants were supported by a male and female therapy team (Mithoefer et al. 2011). The therapeutic model described in the Treatment Manual was based upon initial work with classic psychedelics (Grof 2001;Mithoefer 2017) and early reports of MDMA in a therapeutic setting (Greer and Tolbert 1986). Four of these MAPS-sponsored studies have been published (Mithoefer et al. 2011(Mithoefer et al. , 2013Oehen et al. 2013;Ot'alora et al. 2018), and all six studies demonstrated acceptable safety and promising efficacy results. The MDMA doses selected for phase 2 trials (control-0 mg, 25 mg, 30 mg, 40 mg; active-75 mg, 100 mg, 125 mg) were based on tolerability and subjective effects reported in several prior phase 1 studies (Cami et al. 2000;de la Torre et al. 2000;Grob et al. 1998;Harris et al. 2002;Liechti et al. 2001). Low doses (25 mg, 30 mg, 40 mg) produce some changes in subjective effects that could presumably enhance blinding as an active placebo but would be inadequate for a therapeutic response (Harris et al. 2002). The FDA, after reviewing all available data in 2016, granted Breakthrough Therapy Designation in 2017 and approved the designs of two phase 3 trials that started in 2018.
To optimize the design of the phase 3 trials, we pooled data from six phase 2 trials that had similar study objectives and designs. We aimed to determine how many MDMA sessions are needed to achieve a clinically significant response, what demographic and other baseline variables might impact outcomes, which safety parameters are essential, the optimal dose, and how best to minimize bias and enhance blinding. To that end, the aim of this paper is to present pooled data from randomized clinical trials at different study sites that evaluated the efficacy and safety profile of MDMA-assisted psychotherapy among individuals with PTSD from a range of causes.
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.

Participants
Participants were recruited through internet advertisements, referrals by health professionals, and by word of mouth. Candidates had chronic PTSD with symptoms lasting longer than 6 months and Clinician-Administered PTSD Scale for DSM-IV (CAPS-IV) scores ≥ 50 (all studies except MP-4) or ≥ 60 (MP-4) upon enrollment (see eTable 1 for individual study criteria). Studies enrolled men and women, including civilians and veterans/first responders, aged 18 and older with previous inadequate response to at least one pharmacotherapy and/or psychotherapy. An inadequate response to previous treatment was concluded if participants had a CAPS-IV total score indicating moderate to extreme PTSD at screening. Participants underwent extensive screening by independent examiners, including psychological assessments, physical examinations, laboratory testing, and ECG to identify any possible contraindications to receiving MDMA. The Structured Clinical Interview for DSM-IV Axis I Disorders-Research Version (SCID-I-RV) or the SCID-II was used during screening to detect comorbid disorders, and medical and therapy records from outside providers were reviewed. Participants were not excluded for meeting criteria for anxiety disorders or depression but were excluded if they met criteria for past or current psychotic disorder or Bipolar Disorder 1, or for current borderline personality disorder, or eating disorder with active purging. Other exclusion criteria included significant medical diagnoses (contraindications for MDMA), pregnancy or lactation, and weight under 48 kg. Cardiovascular or cerebrovascular disease was excluded, except in one study where candidates with well-controlled hypertension and no other evidence of vascular disease could enroll after additional screening with nuclear stress test and carotid ultrasound. All therapists maintained a Basic Life Support certification, and a study physician was available by telephone throughout the study. In order to be enrolled, individuals had to meet all inclusion/ exclusion criteria and agree to comply with all planned study visits. All participants confirmed comprehension of study procedures and gave written informed consent.
Participants could not have a diagnosis of substance abuse disorders within 60 days of screening for five studies and within 6 months for one study. Psychiatric medications were tapered and discontinued prior to commencing experimental sessions. Anxiolytics and sedative hypnotics were used asneeded between experimental sessions.

Protocols and treatments
After screening and enrollment, participants were randomized through a web-based system (MP-8, MP-12) or a list generated by a blinded randomization monitor (MP-1, MP-2, MP-4, MP-9) to receive blinded doses of placebo/control (0 mg placebo; 25 mg, 30 mg, or 40 mg MDMA) or active doses of MDMA (75 mg, 100 mg, or 125 mg MDMA) at approximately 1:2 ratio. Doses were administered during two 8-h psychotherapy sessions spaced 3-5 weeks apart. The initial dose was followed approximately 1.5-2.5 h later by an optional supplemental dose equal to half the initial dose. Participants could accept or decline the supplemental dose, and could discuss the choice with the therapy team. The team could withhold the supplemental dose if there were contraindicating circumstances. Participants underwent two to three non-drug 90min therapy sessions prior to the first experimental session. Fifty participants who received 100 mg or 125 mg had a third experimental session, either open label or blinded depending on the study, and one 75 mg participant had a blinded third session before a protocol amendment changed the crossover to occur after two sessions. The control groups subsequently had the option to receive two to three open-label sessions with active dose MDMA in a crossover segment (data not shown). MDMA was synthesized by David Nichols at Purdue University. Gelatin capsules were compounded with lactose to produce equivalent-weight capsules across dose groups.
The same male/female therapy team was present for all therapy sessions for a given participant. There were 18 therapy teams across the six studies. All but one team (MP-2) were trained in the MAPS Therapy Training Program based on the method described in the MDMA-assisted Psychotherapy Treatment Manual (Mithoefer 2017). The method includes periods of introspection alternating with periods of communication between therapists and the participant. The method is aimed at allowing participants to revisit traumatic experiences while staying emotionally engaged even during intense feelings of anxiety, pain, or grief without feeling overwhelmed. The relatively non-directive approach is intended to allow for processing of other psychological, interpersonal, or behavioral aspects of the participants' lives that are likely to arise spontaneously in addition to processing the traumatic memories that led to PTSD.
Experimental sessions took place in a designated area that contained a futon or sofa, as well as artwork or other objects intended to make the space esthetically pleasing. Participants had the option of wearing eye shades and listening to mostly instrumental music during the parts of experimental sessions when they were focused inward. After the 8-h experimental sessions, participants remained at the study site overnight with a supportive attendant. On the following day, they met with the therapists in a 90-min integration session to address and process material that arose during the experimental session. Two to three more integration sessions occurred during the month after each experimental session. For 7 days following each experimental session, the therapy team checked in with the participants in brief telephone calls to assess wellbeing and safety.

Assessments
Assessments were administered at baseline and at follow-up visits occurring 1 to 2 months after the second and third experimental sessions and at additional time points in some studies. Blinded independent raters not present during therapy sessions administered the CAPS-IV. Safety data were collected throughout treatment. Here, we present a limited set of assessments to support the rationale of this paper.

Primary outcome
The CAPS-IV is a semi-structured interview addressing PTSD symptom clusters (re-experiencing, avoidance, negative mood or cognition, and increased arousal) as recognized by DSM-IV (Blake et al. 1995;Nagy et al. 1993;Weathers et al. 2001). The CAPS-IV contains frequency and intensity scores for each of the three symptom clusters that are summed to produce a total severity score, the primary outcome for these studies. The CAPS-IV has a dichotomous diagnostic score assigned on the basis of meeting PTSD diagnostic criteria.

Secondary outcome
The Beck Depression Inventory-II (BDI-II), an established 21-item measure of self-reported depression symptoms (Beck et al. 1996), was administered in four of the six studies (MP-4, MP-8, MP-9, and MP-12). Responses are made on a fourpoint Likert scale and summed to produce an overall score.

Safety outcomes
Safety was assessed by tracking the rates of spontaneously reported reactions (subset of adverse events (AEs) that could be expected based on findings from published studies in healthy volunteers) during experimental sessions and 7 days following, and by recording treatment-emergent AEs (TEAEs), which were not collected on the spontaneously reported reactions list, or were reactions that continued for 7 days or more after experimental sessions. Blood pressure and heart rate were measured in intervals of 15 to 30 min, and body temperature every 60 to 90 min during experimental sessions. Suicidal ideation and behavior were collected at all visits and twice during the 7 days of contact in four of the six studies (MP-4, MP-8, MP-9, and MP-12) using the clinicianadministered Columbia Suicide Severity Rating Scale (C-SSRS) (Posner et al. 2007(Posner et al. , 2011, a structured interview addressing presence and intensity of suicidal ideation and behavior. Participants completed the Paced Auditory Serial Addition Task (PASAT) (Gronwall 1977) and Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) (Randolph 1998) at baseline and 2-month follow-up to determine whether changes in cognitive function had occurred after two sessions with placebo dose or active dose MDMA in specific studies (MP-1, MP-4, and MP-12).

Statistical analysis
Data were pooled across the six studies. Participants who received MDMA (75, 100, 125 mg) were combined into an active dose group; participants who received MDMA (0, 25, 30, 40 mg) were combined for the control group. The modified intent-to-treat set included randomized participants who completed at least one blinded experimental session and a post-baseline assessment. Missing data were not imputed. The safety set included all participants exposed to at least one dose of study drug or placebo.
Group differences in baseline characteristics and demographics were evaluated with χ 2 tests or independent-samples t tests. The primary efficacy evaluation was made with a mixed-effect repeated measure model (MMRM) on change in CAPS-IV total score from baseline to post second experimental session endpoint, and the post third experimental endpoint. The base model included treatment (active/control), baseline CAPS-IV score, and study as a fixed effect, and participant was specified as a random effect. To assess the relationship between outcome measures and age, PTSD duration, sex, race, and prior self-reported Becstasy^use (substances assumed to contain MDMA), these variables were added to the base model one at a time. BDI-II scores were analyzed the same way. AEs were categorized with the Medical Dictionary for Regulatory Activities (MedDRA) in System Organ Classes and preferred terms. AEs, reactions, and suicidal ideation and behavior were summarized descriptively. Independent-samples t tests compared peak vital signs during experimental sessions between groups. Between-group effect size was calculated with Cohen's d (Kadel and Kip 2012). SAS software version 9.3 (SAS Institute Inc., Cary, NC) was used for analyses.

Results
Sample eFigure 1 illustrates flow of participants for these studies. From the 488 telephone-screened, 105 were enrolled and randomized [mean (SD) age, 40.5 (10.5); the majority were white/Caucasian participants (87.6%); and nearly sex balanced (females 58.1%)]. Table 1 displays the characteristics of the sample. Demographic characteristics were approximately matched between treatment arms, and no significant differences were found between groups for demographic and baseline characteristics presented in Table 1. The mean (SD) duration of PTSD was 215.3 (190.3) months, with trauma from various causes. Many participants had a lifetime history of positive suicidal ideation (86.8%) and/or behavior (30.9%). The optional supplemental dose was taken in 179/197 (90.9%) of blinded experimental sessions. The dropout rate was 7.6% (8/105), with six participants terminating early, but having completed at least one experimental session and follow-up assessment.

Primary outcome
The change in CAPS-IV total score (Fig. 1) from Baseline to after the second experimental session was significantly different [t(95) = − 4.25, P < 0.0001] between control (0-40 mg) and active (75-125 mg) groups ( Table 2). The active group had the greatest estimated mean (SE) drop in scores − 30.4 (3.20) compared to the control group − 10.5 (4.46). The between-group Cohen's d effect size was 0.8, indicating a large treatment effect. Study, age, PTSD duration, sex, race, and prior Becstasy^use did not predict outcome in this model.

Secondary outcomes
According to CAPS-IV assessment at the endpoint 1-2 months post two experimental sessions (Table 2), more participants in the active group (54.2%) did not meet PTSD diagnostic criteria than the control group (22.6%). Depression symptom improvement on the BDI-II was greatest for the active dose group, estimated mean (SE) change active − 12.4 (1.84) versus control group − 6.5 (2.69), with the difference between groups trending toward significance [t(61) = − 1.97, P = 0.053].
Depending on the study, after two blinded experimental sessions, most participants in the active dose group had one additional open-label (MDMA 100-125 mg, n = 42) or blinded session (MDMA 75-125 mg, n = 9). The estimated mean change (SE) from baseline to post third session on CAPS-IV for the active dose group was − 45.4 (3.61) with a significant further decline from second to third session [t(95) = − 12.58, P < 0.0001]. The within-participant pre-test (baseline) to post-test Cohen's d effect size increased from 1.4 (post two sessions) to 1.9 (post three sessions). Due to the crossover, there is no between-group comparison for the post third session time point.

Safety and tolerability
Treatment-emergent adverse events (TEAEs) during the blinded treatment segment most commonly reported across all doses included events in the following MedDRA System Organ Classes (SOC): psychiatric disorders, gastrointestinal disorders, and general disorders (eTable 2). The most frequently reported psychiatric TEAEs (Table 3) were anxiety, depressed mood, irritability, and panic attack. On the day of blinded experimental sessions, reactions reported by ≥ 40% of participants in either group were anxiety, dizziness, fatigue, headache, jaw clenching/tight jaw, lack of appetite, and nausea. The majority of expected reactions were rated mild or moderate, and the frequency of reports decreased over the 7 days following an experimental session (eTables 5 and 6). No changes in neurocognitive function were detected (eTable 4).
There were no unexpected MDMA-related SAEs. Four SAEs were reported during the blinded treatment period, including one instance of suicidal ideation (30 mg) ; one SAE of exacerbation of ventricular extrasystoles was reported during an open-label session (125 mg) ) and one SAE of suicidal behavior prior to MDMA exposure in the first experimental session.
There was no suicidal behavior during the treatment period after dosing (eTable 3). At baseline, prior to any drug dosing, the active dose group (46%) had much higher rates of positive suicidal ideation than the control group (16.7%), but the lifetime reports (Table 1) were similar between groups. During the treatment phase, suicidal ideation transiently increased in some participants and was more common in the active MDMA group (eTable 3), although the causal relationship to the psychotherapeutic processing of traumatic memories or to MDMA itself, or to random group differences could not be determined.

Discussion
By pooling data across six phase 2 trials, we found significant symptom reductions in a large sample of participants with PTSD treated with active doses of MDMA combined with psychotherapy. The results informed the design of two phase 3 trials (one now ongoing the other to follow) that were approved through a Special Protocol Assessment by the FDA. The reproducible findings attained by various therapy teams in participants with PTSD arising from different types of traumatic experiences demonstrate the generalizability of this manualized drug-therapy approach and the applicability of the MAPS MDMA Therapy Training BMI, body mass index; CPT, cognitive processing therapy; IPT, interpersonal therapy; CBT, cognitive-behavioral therapy; EMDR, eye movement desensitization and reprocessing therapy; PE, prolonged exposure therapy; C-SSRS, Columbia-Suicide Severity Rating Scale; CAPS-IV, Clinician-Administered PTSD Scale; BDI-II, Beck Depression Inventory a There were no significant group differences (χ 2 or independent t tests) for any variables presented in this table b Sertraline and paroxetine are the only two FDA-approved medications for PTSD. Participants took many other medications for symptom management pre-study that are not presented here. Twelve participants took both sertraline and paroxetine c Lifetime accounts for all suicidal ideation and behavior prior to study, according to participant recall and medical records. According to the C-SSRS scoring guide, scores of four or five on the suicidal ideation category are considered serious ideation, and scores of one or greater are considered positive behavior or ideation. Four phase 2 studies administered the C-SSRS (control group n = 18 and active MDMA group n = 50) d For BDI-II, active group (n = 50) and control group (n = 18) Program. Overall, the treatment was safe and efficacious for civilians and veterans/first responders with chronic PTSD who had previously failed to respond to pharmacotherapies and/or psychotherapy. More than half of the participants had previously undergone first-line trauma-focused psychotherapies, and all but two participants had received some type of psychotherapy prior to study enrollment. MDMA-assisted psychotherapy was effective for these individuals, suggesting a different mechanism of action for MDMA for reducing PTSD symptoms. The data show that when using the Bgold standard^measure of PTSD (CAPS-IV) as a primary outcome measure, with blinded raters, for participants with highly refractory PTSD (mean duration 215.3 months), there was a significant effect after two blinded active doses of MDMA adjunctive with psychotherapy versus psychotherapy with control doses. Notably, more participants in the active dose group (54.2%) no longer met PTSD diagnostic criteria compared to the control group (22.6%). The between-group effect size was large with Cohen's d equal to 0.8. The effect size was used in power calculations for phase 3 trials. Planned enrollment is 100 participants in each phase 3 trial, with an interim analysis and option for sample size adjustment after 60% of participants have completed the primary endpoint. In addition, depression symptoms trended toward greater improvement in participants receiving active MDMA compared with the control group.
After a third experimental session, symptoms on average improved further for the active dose group. The interpretation is limited because the third session was open label for most participants, and there was no control group for comparison due to the open-label crossover after two blinded sessions for most participants. However, it appears that while many people respond adequately after two MDMA sessions, an additional session leads to more participants reaching clinically significant symptom reductions and greater drops in CAPS-IV scores. For this reason, phase 3 trials will include three blinded experimental sessions to maximize response at the primary endpoint (2 months post third experimental session, i.e., 18 weeks post baseline). Phase 2 data showed that 75 mg MDMA produced significant improvement ), yet the sample was quite small (n = 7); therefore, we do not know what the optimal dose is, 75 mg, 100 mg, or 125 mg. There is individual variation in subjective effects of MDMA, and fixed-dose regimens do not account for differences in body weight. To gather more information about optimal dosing, phase 3 trials from 15 sites in the USA, Canada, and Israel will employ a flexible dose regimen. Participants will be randomized to receive equal-weight blinded capsules of inactive placebo or MDMA (80 mg) plus supplemental half-dose unless contraindicated in experimental session one, and then have the choice to escalate the dose to 120 mg with optional supplemental dose (or stay at 80 mg) in the next two sessions. An inactive placebo plus the same psychotherapy will be used as the control group, with the same option to escalate the dose. To minimize bias, a blinded independent rater (IR) pool will administer the primary outcome measure (CAPS-5) to participants across all sites based on availability of IRs. Consecutive assignments to the same IR will not be permitted. Independent raters will remain blinded to the number or timing of CAPS measurements in the study; therefore, we cannot reveal this information until the trials are complete.
The safety and tolerability of limited doses of MDMA in highly controlled therapeutic settings in a PTSD population was adequate, consistent with previous phase 1 studies. There was a dose effect for mean increase in vital signs during MDMA sessions (eTable 7), with values returning or trending toward baseline by the end of the 8-h session. Because vital sign increases did not reach clinically concerning ranges, the  frequency of required vital measurements will be reduced in phase 3 trials to baseline, pre-supplemental dose, and session end. Vital signs at the pre-supplemental dose reading will be taken into consideration before administering the additional half-dose. Neurocognitive measures will not be employed during phase 3 because phase 2 studies showed no evidence of cognitive impairment after two doses of MDMA (eTable 4). MDMA at all doses tested was well tolerated, as demonstrated by the low rate of TEAEs and expected reactions. Most were mild to moderate, resolving as MDMA effects dissipated or during the week following (eTables 5 and 6). During experimental sessions, the active MDMA group had higher incidences of some reactions, including anxiety, dizziness, jaw clenching/tight jaw, lack of appetite, and nausea. Whether reactions are due to the pharmacological effects of MDMA or from augmented trauma processing catalyzed by MDMA effects cannot be determined from the data collected in these studies, but phase 1 studies in healthy individuals report similar reactions to MDMA. During the 7 days following experimental sessions, some reactions occurred more often in the active dose group for the first few days before declining by the end of the week. For this reason, the number of telephone contacts after an experimental session will be less frequent for phase 3 trials, which will require four telephone contacts over 7 days. In accordance with FDA guidance for all psychiatric drugs under development, the C-SSRS will be given at each in-person visit. In phase 2 trials, there were no related deaths or incidence of suicidal behavior after MDMA. The low dropout rate (7.6%) in MDMA trials compared to other PTSD treatments (approximately 17-36%) (Bradley et al. 2005;Marshall et al. 2001;Steenkamp et al. 2015) could be related to the propensity of MDMA to make trauma processing more tolerable with rapid symptom improvements in the days and weeks following. Participants in the placebo/ control group had the opportunity to cross over to receive three open-label (100-125 mg) sessions of MDMA-assisted psychotherapy if they complete the blinded segment which likely motivated participants to complete treatment.
MDMA in the context of psychotherapy was found to have a low potential for abuse. There were no AEs or treatment discontinuation related to Becstasy^seeking or craving, and no reports of use outside the study through the post third session endpoint. Indeed, many participates anecdotally reported that the experimental sessions were not particularly pleasurable experiences, but rather difficult therapeutic work delving into their traumatic memories. Overall, safety outcomes were favorable for use of MDMA in individuals with PTSD in a supportive environment with trained mental health professionals.

Limitations
There are limitations of these trials and the associated pooled data analyses. The sample was nearly gender balanced, but participants and therapists were predominantly White/ Caucasian. Phase 3 studies will evaluate the generalizability to individuals from more diverse ethnic and cultural backgrounds. Across the six trials, there were variations in study design, such as differences in timing of outcome measures, doses tested, number of blinded experimental sessions, and number of participants in each dose group. Drawbacks of pooled data analyses are that multiple doses tested were combined into two groups-control group and active dose group-and that the third experimental session was blinded or open-label full-dose MDMA, depending on the study. Also, there was no control group for a between-group comparison of the post third session; therefore, response after three sessions was limited to a within-subject analysis. Due to small sample sizes, reliability of effect size estimates from individual studies is unknown. Blinding of treatment assignment for psychoactive substances is a recognized challenge. Both psychological and vital sign changes during experimental sessions can be clues to the group assignment. To reduce bias, blinded independent raters who were not present during therapy sessions administered the CAPS-IV. However, participants and therapists often, but not always, accurately guessed dose assignment (Mithoefer et al. 2011Oehen et al. 2013;Ot'alora et al. 2018)-a recognized limitation in clinical trials of all TEAE, treatment-emergent adverse event a Frequency of subjects who reported an expected, spontaneously reported reaction collected during blinded experimental sessions 1 and 2 (only reactions reported by ≥ 40% of participants in any group are displayed; see supplemental for full list of reactions) b Frequency of subjects who self-reported psychiatric adverse events after first drug administration until the day before experimental session 3 (only AEs reported by three or more subjects in either group displayed) drugs with perceivable effects and in all psychotherapy studies where there is no possibility of effective blinding.

Conclusions
Based on the promising safety and efficacy results from these six phase 2 trials, we have designed multi-site, placebo-controlled phase 3 trials that started in late 2018 to evaluate MDMA-assisted psychotherapy in approximately 200 participants with PTSD. Limitations discussed here will be addressed, and if findings are significant and no new safety concerns arise, MDMA could become an FDA-approved treatment for PTSD in the context of psychotherapy by 2021.