The Effects of MDMA-Assisted Therapy on Alcohol and Substance Use in a Phase 3 Trial for Treatment of Severe PTSD

PTSD is frequently comorbid with alcohol and substance use disorders (ASUD), with prior research estimating co-occurrence between 17% and 46%. This comorbidity is commonly attributed to self-medication, since PTSD symptoms often precede substance use, though bidirectional and more nuanced interactions have also been proposed. MDMA-assisted therapy (MDMA-AT) was recently shown in a multi-site, randomized, placebo-controlled Phase 3 trial to reduce severe PTSD symptoms substantially, with 88% showing a clinically significant response and 67% in the MDMA-AT arm no longer meeting diagnostic criteria for PTSD by study end; measures of functioning and depression also improved in that trial, suggesting potential downstream effects on comorbid conditions. Nicholas and colleagues set out to examine whether MDMA-AT for severe PTSD is associated with changes in alcohol and drug use. The present analysis uses data from that Phase 3 trial to compare changes from baseline to study termination on standard screening instruments for hazardous alcohol use (AUDIT) and drug use (DUDIT) in participants randomised to MDMA-AT versus placebo plus therapy, with an emphasis on whether MDMA-AT might reduce or increase alcohol/substance use in this population. The analysis is exploratory and focused on screening-level changes rather than diagnostic outcomes.

This analysis used data from a two-arm, randomized, double-blind, placebo-controlled Phase 3 trial conducted at 15 sites in the US, Canada and Israel. Adults with severe PTSD were screened and enrolled if they had symptoms for at least six months and a CAPS-5 (Clinician-Administered PTSD Scale for DSM-5) Total Severity Score ≥ 35. The Mini-International Neuropsychiatric Interview (MINI) assessed psychiatric disorders and ASUDs in the prior 12 months. Participants could have current mild or early-remission moderate alcohol or cannabis use disorder, but any other active substance use disorder within the prior 12 months led to exclusion; past ASUDs older than 12 months were permitted. After medication taper where applicable, participants received three 90-minute preparatory therapy sessions and were then randomised to receive either MDMA-AT or placebo plus the same therapy. Each participant underwent three 8-hour experimental sessions spaced over the trial; MDMA dosing was a divided dose (initial plus supplemental 1.5–2 hours later): 80 + 40 mg in the first session, escalating to 120 + 60 mg in the second and third sessions unless supplemental dosing was withheld for tolerability or participant preference. Each experimental session was followed by three 90-minute integration therapy sessions over 3–4 weeks. Blinding included participants, site staff and the sponsor until database lock. Primary alcohol and drug-related outcomes for this analysis were the Alcohol Use Disorders Identification Test (AUDIT), a 10-item self-report screening tool for hazardous alcohol consumption over the prior 12 months, and the Drug Use Disorders Identification Test (DUDIT), an 11-item parallel instrument assessing patterns of drug use and related problems. In this analysis AUDIT and DUDIT were collected at baseline (standard 12-month lookback) and at study termination with instructions to report since end of treatment. Statistical analyses used descriptive statistics, t-tests and chi-square tests for group comparisons, and models comparing change scores between treatment arms; both unadjusted and models adjusted for baseline values were reported. Hedge's g effect sizes were calculated for statistically significant effects. Analyses were conducted in SAS 9.4. The extracted text indicates that participants with complete baseline and termination AUDIT/DUDIT data (82 of 90 randomised) were included in the present analysis.

Ninety participants were randomised and received study treatment; three in the MDMA arm and four in the placebo arm withdrew, leaving 82 participants with complete baseline and termination AUDIT/DUDIT data used in these analyses. The analysed sample was predominantly female (64.6%), White (80.3%), non-Hispanic or Latino (92.7%), with mean age 41.4 (SD 12.22). Baseline clinical measures indicated severe PTSD (mean CAPS-5 43.8 (6.00)), severe depression (mean BDI-II 32.3 (13.01)), and significant functional impairment on the adapted Sheehan Disability Scale (mean 7.06 (1.89)). At baseline, 21 (25.61%) reported past alcohol use disorder only, 14 (17.07%) reported past substance use disorders, and 2 placebo participants reported current mild cannabis use disorder. Overall, 69 (84.2%) had AUDIT ≥ 1 and 48 (58.5%) had DUDIT ≥ 1 at baseline, and there were no statistically significant baseline differences between treatment arms on demographic variables or AUDIT/DUDIT scores. With respect to primary comparisons, MDMA-AT was associated with a greater reduction in AUDIT scores from baseline to study termination than Placebo+Therapy: mean change Δ = -1.02 (SD 3.52) for MDMA-AT versus Δ = 0.40 (SD 2.70) for placebo, yielding F(80, 1) = 4.20, p = 0.0436, and Hedge's g = 0.45 (reported as the effect size). This effect corresponds to a small-to-moderate standardised effect size. However, when models adjusted for baseline AUDIT values the between-group difference in AUDIT change scores was no longer statistically significant (p = 0.52). There were no significant between-group differences in DUDIT change scores. Correlational analyses found no significant linear relationship between AUDIT change scores and baseline or change scores for CAPS-5, BDI-II, or SDS in the overall sample. Within the MDMA-AT group only, AUDIT change scores were positively correlated with baseline CAPS-5 (r = 0.32, p = 0.04), suggesting greater baseline PTSD severity was associated with larger reductions in AUDIT. The extracted text also states that there was no evidence of increased substance use, abuse, dependence, or increased MDMA use during and two months following treatment, though detailed incidence data are not provided in the extracted text.

Nicholas and colleagues interpret these exploratory findings as suggesting that MDMA-AT for severe PTSD was associated with a statistically significant decrease in hazardous alcohol use as measured by AUDIT change scores, and that the MDMA-AT protocol did not increase risk of illicit drug use in the short term. The authors caution that the AUDIT effect was no longer significant after adjustment for baseline scores, and note there were no baseline differences between groups, indicating the result should be considered preliminary. The investigators emphasise key limitations that constrain interpretation: the trial enrolled few or no participants with current diagnosable alcohol use disorder despite allowing mild cases, and other active substance use disorders were excluded, producing a restricted distribution of AUDIT and DUDIT scores that limits sensitivity to detect change or to examine diagnostic cut-offs and factor-level changes (consumption, consequences, dependence). Heterogeneity in types of substances used and low baseline severity of drug use may explain the lack of a detectable DUDIT effect. The authors also describe the exploratory nature of the analysis. In terms of implications, the discussion notes that long-term follow-up is under way and will report AUDIT/DUDIT changes and self-reported MDMA craving and use at least 12 months post-trial. The authors position these results alongside previous Phase 2 follow-up data that suggested limited abuse potential of MDMA when administered in controlled clinical settings with careful eligibility screening and therapeutic support. They suggest that, given the observed reductions in alcohol-related screening scores and prior open-label findings in primary alcohol use disorder, further investigation of MDMA-AT as an integrated treatment for co-occurring PTSD and alcohol use disorder, or for alcohol use disorder alone, may be warranted.