The safety and efficacy of ±3,4-methylenedioxymethamphetamine-assisted psychotherapy in subjects with chronic, treatment-resistant posttraumatic stress disorder: the first randomized controlled pilot study

Posttraumatic stress disorder (PTSD) is described as a chronic, debilitating anxiety disorder with high prevalence and limited treatment effectiveness. Mithoefer and colleagues note that existing pharmacotherapies (notably SSRIs) and psychotherapies produce meaningful benefit for many patients but leave 25%–50% of trial participants with inadequate response; combined pharmacotherapy and psychotherapy data are limited and mixed. Given the substantial unmet need, the authors introduce MDMA as a pharmacological adjunct that was historically used to facilitate psychotherapy and which, in modern Phase I work, has demonstrated an acute subjective profile of reduced fear, increased sociability and preserved clarity of consciousness. This pilot Phase II study set out to test two linked hypotheses: that MDMA could be administered without harm to carefully screened patients with chronic, treatment-resistant PTSD, and that when paired with manualised psychotherapy it would produce greater reductions in PTSD symptom severity than the same psychotherapy with inactive placebo. The trial therefore evaluates safety, tolerability and preliminary efficacy of MDMA-assisted psychotherapy in this population, and gathers feasibility data for future trials.

This was a randomised, double-blind, placebo-controlled pilot trial (Stage 1) with an open-label crossover option for placebo subjects (Stage 2). Twenty adults aged 21–70 with chronic, treatment-resistant PTSD were enrolled between March 2004 and January 2008 and followed until September 2008. Participants were required to meet DSM-IV-R criteria for crime- or war-related chronic PTSD, have a Clinician-Administered PTSD Scale (CAPS) score of ≥50 despite at least 3 months of prior SSRI or SNRI treatment and at least 6 months of psychotherapy, and to be free of major medical conditions and certain psychiatric exclusions (for example current Borderline Personality Disorder was excluded). Urine drug screening and a battery of medical and laboratory evaluations were performed at screening. Allocation used a 60:40 randomisation ratio to MDMA (n = 12) versus inactive placebo (lactose; n = 8), with replacement of early dropouts to preserve group ratios. The blind was preserved for subjects, investigators and an independent rater who conducted outcome assessments; it was broken for each subject after the 2-month follow-up visit. The core experimental intervention comprised two 8–10 h psychotherapy sessions in which participants received either MDMA (first dose 125 mg, optional supplemental half-dose 62.5 mg given 2–2.5 h later for later enrollees) or placebo, plus preparatory and integration non-drug psychotherapy sessions. Sessions took place in an outpatient clinic with overnight stay capability, a male–female co-therapist team, music-guided introspective periods and flexible therapeutic discussion. Subjects were required to taper and abstain from psychotropic medications during participation except for as-needed sedative hypnotics or anxiolytics ('rescue medications'). Primary outcome was CAPS total severity score measured at baseline, about 4 days after each experimental session and at 2 months after the second session. Secondary measures included the Impact of Events Scale–Revised (IES-R) and the SCL-90-R. Neurocognitive testing (RBANS, PASAT, RCFT) and physiological monitoring (blood pressure, pulse, temperature) were included. Per-protocol statistical analysis used repeated-measures ANOVA to test group-by-time differences, with Holm’s sequential Bonferroni correction applied to post-hoc comparisons.

Twenty subjects entered the trial; 15 had multiple prior medication trials (mean 4.2 different psychiatric drugs) and 15 had completed more than one course of psychotherapy. Mean duration of PTSD was reported as about 19 years. Two subjects (both randomised to MDMA) dropped out before the second experimental session; their large CAPS reductions after one MDMA session were noted but excluded from the primary analysis. Baseline characteristics were similar between groups except for duration of prior therapy; that covariate did not affect CAPS results when tested. Physiological and acute subjective effects followed the expected MDMA time-course: onset 45–75 minutes after dosing, peak at 2–2.5 hours, and duration of about 4–6 hours depending on whether a supplemental dose was given. Blood pressure, pulse and temperature rose more in MDMA sessions than in placebo sessions but returned to baseline by 6 hours; no medical complications or pharmacological interventions were required. Reported acute side effects more common with MDMA on session days included jaw tightness, nausea, feeling cold, dizziness, loss of appetite and balance impairment, whereas anxiety, insomnia, headache and fatigue were reported as equally or more common in the placebo group on the day of sessions. In the week after sessions, both groups reported fatigue, anxiety, low mood, headache and nausea at similar rates. No serious drug-related adverse events occurred, and neurocognitive testing showed no evidence of impairment attributable to MDMA. On the primary outcome, CAPS scores improved over time in both arms (Time effect F(3,17) = 40.292, p < 0.0005), but the MDMA group showed significantly greater improvement (Time*Group interaction F(1,17) = 7.173, p = 0.015). Post-hoc comparisons with Holm correction indicated non-significant difference at the first post-baseline time point (Time 1 p = 0.966) but significant group differences at subsequent time points (e.g. Time 2 p = 0.013; Time 3 p = 0.002; Time 4 p = 0.013 under the study’s adjusted alpha levels). The IES-R showed a comparable pattern: overall improvement (Time F(3,17) = 11.003, p < 0.0005) with a significant Time*Group interaction (F(1,17) = 3.290, p = 0.027) and significant post-hoc differences at later time points. Clinical response, defined as >30% reduction in CAPS total severity, occurred in 83.3% (10/12) of MDMA-treated subjects versus 25% (2/8) of placebo subjects during Stage 1. Ten MDMA subjects no longer met DSM-IV criteria for PTSD at 2 months versus two in the placebo group. In the open-label Stage 2, seven subjects received MDMA and all seven met the clinical response criterion. The between-group effect size reported in the discussion was 1.24. Additional integration psychotherapy sessions were provided more often after MDMA sessions (20 sessions to seven of 13 subjects) than after placebo (one additional session); the authors note this may confound interpretation but argue it does not explain early improvements seen within days of the first MDMA session. Supplemental dosing (per protocol amendment) did not show a significant effect: comparison of four subjects who received a supplemental dose versus eight who did not yielded no significant Group by Time difference (ANOVA F(1,9) = 0.413, p = 0.637). Blinding was largely ineffective: 19/20 participants (95%) correctly guessed their assignment and therapists guessed correctly in all cases (three with uncertainty). Rescue medications (zolpidem, benzodiazepines) were used by many participants but did not differ significantly between conditions (zolpidem after 60.7% of MDMA sessions versus 68.8% of placebo sessions, p = 0.77; benzodiazepines after 47.0% versus 37.5%, p = 0.57). Decreases in CAPS scores did not differ between subjects who did or did not receive benzodiazepines (p = 0.98).

Mithoefer and colleagues interpret the findings as preliminary evidence that MDMA-assisted psychotherapy can be administered safely to a carefully screened sample of people with chronic, treatment-resistant PTSD and that, when paired with manualised psychotherapy, it produced clinically and statistically significant reductions in PTSD symptoms relative to the same psychotherapy with inactive placebo. The investigators highlight the large between-group effect size (1.24) and the high proportion of responders (83% in the MDMA arm) as indicators of clinical significance, noting that several participants regained work function. The authors position these results as promising relative to existing treatments, particularly given the refractory nature and long illness duration (mean ~19 years) of the sample. They emphasise methodological strengths: prospective, double-blind design in Stage 1, use of the CAPS administered by a blinded independent rater, standardised outcome measurement, enrolment of a treatment-resistant cohort and careful medical monitoring. Several limitations are acknowledged. The trial is small and a Phase II pilot by design; the sample was majority female and entirely Caucasian, limiting generalisability. Blinding was transparently ineffective because the subjective effects of MDMA and associated physiological changes were obvious to participants and therapists, raising the possibility of expectancy effects. The authors note that the placebo group did respond in the open-label MDMA phase, arguing against an intrinsic greater treatment resistance in that group, but concede a placebo effect cannot be ruled out and must be addressed in future trials with improved blinding and dose–response designs. The provision of additional integration psychotherapy sessions more often after MDMA sessions is recognised as a potential confounder, though the investigators argue it does not account for early symptom change seen shortly after the first MDMA session. Follow-up was limited to 2 months post-second session for the main analysis; long-term durability is being studied separately in a 1–5 year follow-up cohort. The absence of formal therapist adherence measures in this pilot is noted, and the authors report development of a treatment manual and adherence metrics for future studies. Practical and implementation considerations are discussed: MDMA-assisted psychotherapy in this model requires extended therapist contact (all-day sessions and overnight stays), careful preparation and integration, and clinic facilities for prolonged sessions, which may raise initial costs but could be cost-effective for a subset of chronic, treatment-resistant patients. The authors conclude that the pilot data justify further research to confirm efficacy, refine therapeutic components, improve methodology (including blinding), and investigate mechanisms of action.