Treating posttraumatic stress disorder with MDMA-assisted psychotherapy: A preliminary meta-analysis and comparison to prolonged exposure therapy

Posttraumatic stress disorder (PTSD) is a chronic and often debilitating condition defined by re-experiencing, avoidance, hyperarousal, and negative mood/cognition changes. Epidemiological estimates cited by the authors put lifetime prevalence in the general population at 6.8%, with substantially higher rates reported among Iraq and Afghanistan veterans. Existing treatments have important limitations: only two drugs (sertraline and paroxetine) are approved, many psychotherapies incur high dropout, and prolonged exposure (PE) therapy—while evidence-based—can be intolerable or inaccessible to many patients. The authors note mechanistic and practical barriers to current treatments, including impaired therapeutic alliance, a narrow therapeutic window for arousal, and avoidance that drives attrition in exposure-based approaches. Against this background, the paper considers MDMA-assisted psychotherapy (MDMA-AP) as an emerging option. MDMA is described as a psychotherapeutic adjunct that may reduce anxiety, increase trust and insight, and widen the “therapeutic threshold,” potentially improving engagement with trauma-focused work. Two clinical MDMA-AP trials had been published at the time. Amoroso and Workman set out to produce a preliminary meta-analysis of those MDMA-AP trials and to compare their cumulative effect sizes and dropout rates with those reported in an existing meta-analysis of PE, arguing that effect size (rather than p-values alone) is a useful comparative metric when sample sizes differ widely.

The investigators performed a targeted literature search using PsycINFO, PubMed and Google Scholar to identify published clinical trials of MDMA-AP and a single, reliable meta-analysis of PE. Search terms for the PE meta-analysis included ‘‘meta,’’ ‘‘exposure,’’ ‘‘PTSD’’ and ‘‘prolonged,’’ while the MDMA-AP search used terms such as ‘‘MDMA,’’ ‘‘double blind,’’ ‘‘placebo controlled,’’ ‘‘posttraumatic stress’’ and ‘‘therapy.’’ Inclusion criteria for MDMA-AP trials required participants to meet DSM-III-R, DSM-IV or DSM-IV-R criteria for PTSD, a randomised double-blind placebo-controlled design (crossover allowed), and sufficient sample size to permit inferential statistics. Three MDMA studies were identified but one was excluded because it had been terminated prematurely, lacked adequate experimental controls and did not report inferential statistics. The two included MDMA trials are described in the extracted text as one randomised, placebo-controlled, double-blind crossover study of 23 participants and a second study of 12 treatment-resistant PTSD patients; the extracted text also reports an aggregate MDMA sample figure inconsistently (see below). Effect sizes were the main comparative metric. The PE meta-analysis reported Hedges' g, while the MDMA trials reported Cohen's d; the authors converted the MDMA estimates to Hedges' g and calculated a weighted average effect size across the MDMA studies. Where secondary outcome results were reported as F-tests rather than effect sizes, the F statistics were converted to Cohen's d prior to the Hedges' g correction. Dropout rates were compared as follows: for the PE meta-analysis the authors calculated the percentage dropout per PE arm from the included trials (only participants assigned to PE were counted), while for the MDMA trials dropouts were divided by the total number of trial participants because those trials used crossover and active placebo designs. Heterogeneity was assessed using the I2 statistic, and publication bias was addressed by calculating a fail-safe N based on mean Z-scores. The extracted text contains some inconsistencies in reported sample totals and ancillary numbers; where a detail was unclear in the extraction, this is noted in the Results and Discussion summaries rather than being inferred.

Primary outcomes: The pooled MDMA-AP effect for clinician-observed PTSD measures was large. The cumulative Hedges' g reported for MDMA-AP primary outcomes was 1.17 (SE=0.09; 95% CI 0.38–1.90; p=0.033). The PE meta-analysis similarly reported a large effect for primary clinician-rated outcomes; while the extracted text does not consistently restate the PE primary estimate in the body, the PE meta-analysis is described as producing a large Hedges' g (the abstract reports PE g=1.08). Secondary outcomes: For patient self-report measures the PE meta-analysis produced a large pooled effect (Hedges' g=0.77; SE=0.12; 95% CI 0.53–1.01; p<0.001). The MDMA-AP pooled secondary outcome effect calculated in this analysis was also large (Hedges' g=0.87; 95% CI 0.01–1.79; p=0.049). The MDMA secondary-effect calculation required converting a reported time×group interaction F-test into an effect-size estimate. Dropout rates and safety: Average dropout across the PE trials included in the meta-analysis was 27.0% (SD=10.8%). The MDMA-AP studies had a lower average dropout of 12.7% (SD=5.6%). The extracted text states that no psychiatric or physical emergencies occurred in the MDMA trials. Reported reasons for MDMA trial dropouts included travel difficulties and resumption of antidepressant medication, with two participants reported to have left because of adverse effects (one of these was in the active placebo group). Heterogeneity and publication bias: The MDMA-AP pooled estimate showed no observed heterogeneity (I2=0%), whereas the PE meta-analysis reported substantial heterogeneity (Cochran's Q(12)=59.90, I2=79.9%). Fail-safe N calculations were presented to assess robustness to unpublished null studies; the extracted text states that 446 null-effect studies would be required to overturn the PE meta-analysis and that 135 such studies would be required to render the MDMA-AP meta-analysis non-significant. The extraction contains fragmented phrases around fail-safe thresholds that could not be fully resolved, so the summary reports the clearer figures provided.

Amoroso and colleagues interpret their findings as indicating that MDMA-AP yields treatment outcomes comparable to, and in some metrics slightly larger than, those reported for PE. Both interventions produced large effect sizes on clinician-rated and self-reported PTSD measures in the comparisons undertaken. The authors highlight several clinical distinctions that might explain differences in acceptability and dropout: MDMA-AP sessions are much longer (typically about eight hours per drug session, with several non-drug follow-ups) and are described as more patient-centred, whereas PE sessions are shorter (around 60 minutes) and can be experienced as rigid and emotionally taxing. These procedural differences are offered as possible reasons MDMA-AP trials showed lower dropout. The discussion acknowledges important limitations. Most prominently, the PE meta-analysis pooled 13 studies with much larger aggregate sample size (n=675) while the MDMA-AP comparison rested on two small trials (the extracted text gives per-trial Ns of 23 and 12, and elsewhere an aggregate MDMA figure appearing as n=37; these totals are inconsistently reported in the extraction). The authors emphasise that effect-size metrics are intended to be relatively robust to sample-size differences, but they nonetheless caution against overinterpretation. Other differences that complicate direct comparison include participant characteristics: the MDMA trials specifically enrolled chronic, treatment-resistant cases, whereas the PE meta-analysis included participants meeting standard DSM criteria without the same treatment-resistance requirement. Trial design differences are noted as well: PE comparisons often used psychological placebo or waitlist controls, while MDMA trials used active placebo conditions in which participants received the same psychotherapy but without an active MDMA dose. The potential confounding influence of concomitant psychotropic medications is also raised; the PE trials included many participants on psychotropics, whereas the MDMA trials required washout of medications prior to dosing. On safety and public-health implications, the authors report that the MDMA-AP trials appeared safe within the trial context and suggest MDMA-AP merits further research given the high burden of PTSD, substantial economic costs and the limits of current treatments. They stop short of asserting superiority and frame MDMA-AP as a promising alternative, particularly for patients who do not tolerate or respond to PE. The authors recommend that emerging treatments such as MDMA-AP be considered and more thoroughly investigated, while acknowledging the methodological and sample-size constraints of the present analysis.