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

Posttraumatic stress disorder (PTSD) is a common, disabling condition associated with intrusive memories, negative changes in mood and cognition, hyperarousal, avoidance and substantial reductions in quality of life. Existing trauma-focused psychotherapies are first-line treatments but are often inaccessible or ineffective for many patients, and only two drugs (sertraline and paroxetine) are FDA-approved for PTSD; pharmacotherapies more broadly fail to produce adequate response in an estimated 40–60% of patients. The resulting unmet need has prompted investigation of novel approaches that combine pharmacological agents with psychotherapy. Mithoefer and colleagues evaluate pooled data from six Phase II randomized, double-blind trials of MDMA-assisted psychotherapy conducted between 2004 and 2017 to inform the design of Phase III trials. Specifically, the investigators sought to define how many MDMA sessions are needed for clinically meaningful benefit, to explore dose and baseline predictors of outcome, to characterise essential safety parameters, and to identify trial-design features that minimise bias and improve blinding across diverse sites and participant groups.

Six randomised, double-blind Phase II studies at five sites (three in the USA, one in Canada, one in Switzerland, one in Israel) were pooled. Participants were adults (aged 18+) with chronic PTSD (symptom duration >6 months) and CAPS-IV scores ≥50 (except one study requiring ≥60). Recruitment included civilians and veterans/first responders, and candidates were required to have had an inadequate response to at least one prior pharmacotherapy and/or psychotherapy. Exclusion criteria included current or past psychotic disorder, Bipolar I, current borderline personality disorder, active purging-type eating disorder, significant medical contraindications to MDMA, pregnancy/lactation and low body weight; substance-abuse disorders were exclusionary within windows specified by study. Psychiatric medications were tapered before experimental sessions and anxiolytics/sedative hypnotics were permitted as needed between sessions. Randomisation employed web-based systems or blinded lists, with an approximate 1:2 allocation to placebo/control-dose MDMA (0, 25, 30, 40 mg) versus active-dose MDMA (75, 100, 125 mg). Experimental treatment consisted of two 8-h manualised psychotherapy sessions spaced 3–5 weeks apart, each with an optional supplemental half-dose given 1.5–2.5 h after the initial dose. Participants received two to three non-drug 90-min preparatory therapy sessions and multiple integration sessions (including an overnight stay after experimental sessions and a 90-min integration the following day); brief safety telephone checks were conducted for 7 days after each session. Depending on the study, some participants received a third experimental session (open-label or blinded) and control participants could cross over to receive active sessions after completing the blinded segment. The primary outcome was change in CAPS-IV total severity score (a structured interview measure of PTSD symptoms) from baseline to the post-second-session endpoint; CAPS-IV diagnostic status (meets/does not meet PTSD criteria) was also recorded. The Beck Depression Inventory-II (BDI-II) was a secondary outcome in four studies. Safety assessments included spontaneously reported reactions during sessions and for 7 days after, treatment-emergent adverse events (TEAEs), vital signs during sessions, C-SSRS assessment of suicidal ideation/behaviour at visits and during some follow-up calls, and cognitive testing (PASAT and RBANS) in selected studies. Independent, blinded raters who did not attend therapy sessions administered CAPS-IV. For analysis, participants receiving 75–125 mg MDMA were combined into an active-dose group and those receiving 0–40 mg into a control group. The modified intent-to-treat set comprised randomised participants who completed at least one blinded experimental session and a post-baseline assessment; missing data were not imputed. Efficacy was evaluated with mixed-effect repeated measures models (MMRM) including treatment, baseline CAPS-IV and study as fixed effects and participant as a random effect. Potential moderators (age, PTSD duration, sex, race, prior self-reported use of substances purported to contain MDMA) were added to the base model one at a time. Between-group effect sizes were reported using Cohen’s d; safety outcomes were summarised descriptively.

Of 488 telephone-screened candidates, 105 were enrolled and randomised. The pooled sample had a mean (SD) age of 40.5 (10.5) years, was predominantly White/Caucasian (87.6%) and nearly sex-balanced (58.1% female). Mean PTSD duration was reported as a long chronic course (extraction shows an unclear figure for months), and many participants had lifetime histories of suicidal ideation (86.8%) and suicidal behaviour (30.9%). The optional supplemental dose was taken in 179 of 197 (90.9%) blinded experimental sessions. Overall dropout was low at 7.6% (8/105). Primary efficacy results showed a statistically significant greater reduction in CAPS-IV total score from baseline to after the second experimental session for the active-dose MDMA group compared with the control-dose group [t(95) = -4.25, P < 0.0001]. The active group’s estimated mean (SE) change was -30.4 (3.20); the extracted text does not clearly report the control group’s corresponding mean change value, but the between-group effect size was reported as Cohen’s d = 0.8, indicating a large effect. Study, age, PTSD duration, sex, race and prior self-reported 'Becstasy' use did not predict outcome in the MMRM. On diagnostic outcome, 54.2% of participants in the active group no longer met CAPS-IV diagnostic criteria for PTSD at the 1–2 month post-second-session endpoint, compared with 22.6% in the control group. Depressive symptoms measured by BDI-II (available in four studies) showed greater improvement in the active group (estimated mean change -12.4 [1.84]) versus the control group (-6.5 [2.69]), with the between-group comparison trending toward statistical significance [t(61) = -1.97, P = 0.053]. For participants who received a third experimental session (most of whom received open-label full-dose MDMA), the active group’s estimated mean change on CAPS-IV from baseline to post-third session was -45.4 (3.61), and there was a significant within-participant further decline from the second to third session [t(95) = -12.58, P < 0.0001]. Because many third sessions were open-label or occurred after crossover, no between-group comparison was available for that time point. Safety and tolerability findings indicated that the most common TEAEs during the blinded treatment period fell into MedDRA System Organ Classes for psychiatric disorders, gastrointestinal disorders and general disorders. Frequently reported psychiatric TEAEs included anxiety, depressed mood, irritability and panic attack. Reactions occurring on the day of experimental sessions in ≥40% of participants in either group included anxiety, dizziness, fatigue, headache, jaw clenching/tight jaw, decreased appetite and nausea; most reactions were rated mild or moderate and declined over the 7 days after sessions. No changes in neurocognitive function were detected in the studies that tested cognition. Four serious adverse events (SAEs) were reported during the blinded treatment period, including one episode of suicidal ideation (at 30 mg), one exacerbation of ventricular extrasystoles during an open-label 125 mg session, and one incident of suicidal behaviour that occurred prior to MDMA exposure in a first experimental session. There were no unexpected MDMA-related deaths, and no suicidal behaviour reported during the treatment period after dosing. Baseline rates of positive suicidal ideation differed between groups (46% active vs 16.7% control), and transient increases in suicidal ideation during treatment were observed in some participants and were more common in the active-dose group; the causal relationship to MDMA, psychotherapeutic processing, or baseline differences could not be determined from the available data. Overall, MDMA at the doses tested was described as well tolerated within the controlled therapeutic setting, with most adverse reactions resolving as acute effects dissipated or over the week following sessions.

Mithoefer and colleagues interpret the pooled Phase II data as demonstrating significant, reproducible reductions in PTSD symptom severity when MDMA is administered as an adjunct to manualised psychotherapy. The findings were consistent across multiple therapy teams, different types of traumatic aetiologies and participant subgroups, which the authors argue supports the generalisability of the MAPS therapy model and the associated therapist training programme. The observed between-group effect size (Cohen’s d = 0.8) and the higher proportion of participants no longer meeting CAPS-IV diagnostic criteria in the active group informed power calculations and the design parameters for two Phase III trials approved by the FDA via Special Protocol Assessment. Results influenced several concrete Phase III design decisions. Because many participants showed additional improvement after a third experimental session, the Phase III primary endpoint will be set at 2 months post-third session (approximately 18 weeks from baseline) and the Phase III protocol will include three blinded experimental sessions. To better characterise optimal dosing, Phase III trials will use a flexible dosing strategy (blinded capsules with an initial 80 mg and optional supplemental half-dose, with the option to escalate to 120 mg in later sessions). Blinding and rater bias will be addressed by employing a blinded independent rater pool for CAPS-5 assessments, reducing the chance of the same rater conducting consecutive assessments and keeping raters blind to timing/number of measurements when possible. In terms of safety and tolerability, the authors note that vital-sign elevations during MDMA sessions were dose-related but did not reach clinically concerning ranges and returned toward baseline by session end; consequently, Phase III trials will reduce the frequency of intra-session vital-sign measurements to baseline, pre-supplemental-dose and session-end, while retaining pre-supplemental-dose checks to guide whether an additional dose is given. Cognitive testing will be omitted from Phase III because Phase II data did not show evidence of cognitive impairment after two doses. Telephone follow-up frequency after sessions will be adjusted (Phase III will require four calls over 7 days) and the C-SSRS will be administered at each in-person visit to monitor suicidal ideation/behaviour. The authors acknowledge important limitations. The pooled sample was predominantly White/Caucasian, limiting ethnic and cultural generalisability. Heterogeneity across the six studies included variations in doses tested, timing of outcome measures, number of blinded experimental sessions and differences in whether the third session was open-label or blinded; these factors complicate pooled analyses and required collapsing multiple dose levels into binary active/control groups. Small sample sizes within individual dose strata reduce the precision of dose-specific effect estimates, and effective blinding was challenging given perceivable psychoactive effects, with participants and therapists often able to guess assignment. The third-session analyses were largely within-subject and open-label for most participants, so between-group conclusions at that time point are not possible. The investigators highlight that these limitations were considered in planning Phase III to improve generalisability, dosing information and mitigation of bias.

The pooled results from six Phase II trials indicated promising efficacy and an acceptable safety profile for MDMA-assisted psychotherapy in adults with chronic, treatment-resistant PTSD. On the basis of these data, the investigators designed multi-site, placebo-controlled Phase III trials enrolling approximately 200 participants that began in late 2018. If Phase III findings are positive and no new safety concerns emerge, the authors state that MDMA-assisted psychotherapy could become an FDA-approved treatment for PTSD within the projected timeline discussed in the paper.