Durability of improvement in post-traumatic stress disorder symptoms and absence of harmful effects or drug dependency after 3,4-methylenedioxymethamphetamine-assisted psychotherapy: a prospective long-term follow-up study

Post-traumatic stress disorder (PTSD) is often chronic and disabling, with many patients remaining symptomatic despite available pharmacological and psychotherapeutic treatments. Prior reports suggested that 3,4-methylenedioxymethamphetamine (MDMA), given in conjunction with psychotherapy, could produce substantial clinical benefits. MDMA is pharmacologically a monoamine releaser with prominent serotonergic action, and it also elevates oxytocin and alters neural responses related to threat and social cognition (for example, reduced amygdala activation and altered recognition of emotional facial expressions). These pharmacological and behavioural effects have been hypothesised to enhance psychotherapy by improving access to painful material, modulating arousal, and strengthening therapeutic alliance. Mithoefer and colleagues report a long-term follow-up (LTFU) of participants from their earlier randomised trial of MDMA-assisted psychotherapy. This study aimed to evaluate durability of PTSD symptom improvement, the presence of any long-term harms (including neurocognitive decline or substance abuse), and participants’ subjective appraisals of benefit, by reassessing participants many months to years after their final MDMA session.

The LTFU builds on an earlier randomised, double-blind trial in which 20 participants with treatment-resistant PTSD were allocated to psychotherapy with active MDMA (n = 12) or to psychotherapy with inactive placebo (n = 8). Each arm included two 8-hour drug-assisted sessions scheduled 3–5 weeks apart, plus weekly non-drug psychotherapy sessions. After the blinded phase, seven of the eight psychotherapy-only participants crossed over to open-label MDMA-assisted psychotherapy, yielding 19 participants who ultimately received MDMA. A protocol amendment allowed the last eight enrollees to receive a third MDMA-assisted session; consequently, 8 subjects had three MDMA sessions and 11 had two prior to the LTFU. All 20 original subjects were invited to the LTFU; the analysis focused on the 19 who had received MDMA. Data collection comprised the Clinician-Administered PTSD Scale (CAPS) as the primary outcome, the Impact of Events Scale–Revised (IES-R) as a secondary outcome, a bespoke LTFU questionnaire addressing perceived benefits/harms, subsequent treatments, illicit drug use, and changes in cognition, and the NEO-PI-R personality inventory (NEO results were not reported here). Questionnaires and the IES-R were mailed to participants; the CAPS was administered in person or by telephone by the same independent rater who had assessed participants during the original trial. The interval from final MDMA session to LTFU assessment varied widely (CAPS/IES-R: 17–74 months, mean 45.4 months; questionnaire: 10–74 months, mean 40.8 months). For analysis, the investigators compared each participant’s 2-month post-treatment scores (the short-term endpoint reported in the original trial) with the LTFU scores. Independent t-tests were used for continuous global scores and Mann–Whitney U tests for ordinal questionnaire outcomes, with descriptive statistics for questionnaire items. An intent-to-treat approach was also applied: participants who did not complete the CAPS at LTFU were treated as negative outcomes in that analysis. A small internal comparison between participants who had received two versus three MDMA sessions found no statistically significant differences on CAPS or IES-R at LTFU, and therefore LTFU scores were compared to the 2-month scores following the second MDMA session.

Of the 19 participants who received MDMA-assisted psychotherapy, 16 completed the CAPS and IES-R at LTFU; all 20 completed the LTFU questionnaire, and data analysis excluded the single subject who had never received MDMA. The investigators report that mean CAPS and IES-R scores at LTFU for the 16 completers did not differ statistically from their 2-month post-treatment scores, indicating maintenance of short-term gains at long-term follow-up. Two CAPS completers (2/16, 13%) had CAPS scores above 50 at LTFU, a threshold indicating moderate-to-severe PTSD and representing relapse above the original study entry cutoff. Using an intent-to-treat approach that treated the three CAPS non-completers as negative outcomes and counted the two relapses, 5 of 19 participants (26%) were classified as having negative outcomes at LTFU; considering only completers, 2 of 16 (13%) had negative outcomes. The bespoke questionnaire indicated that all subjects who received MDMA reported some degree of benefit (median degree of benefit = 5 on a 1–5 scale) and persistence of benefit (median = 5), and none reported being harmed by participation. Comparison between CAPS completers and non-completers showed no significant differences in reported degree or persistence of benefit. Participation in psychotherapy decreased from 16/19 (84%) at study entry to 8/19 (42%) at LTFU; types of therapy reported at LTFU included cognitive behavioural therapy and EMDR among others. Illicit drug use during the follow-up period was reported but appeared limited: cannabis was the most commonly reported substance (n = 8, with frequency from once to occasionally), one participant reported psilocybin use, and one participant reported a single subsequent use of “ecstasy” in a quasi-therapeutic non-clinical setting, which the participant found unsatisfactory and did not repeat. The investigators report no cases of subsequent substance abuse attributable to study participation. On questions about cognition, the paper reports that participants subjectively reported either no change or improvements in cognition, memory and concentration; the extracted text states that seven participants reported no change while 13 reported improvements, although those counts do not arithmetically align with the sample size, so the precise distribution is unclear from the extraction. Participant free-text comments frequently described meaningful benefits (for example, "The therapy made it possible for me to live") alongside the treatment’s demanding nature (for example, "one of the toughest things I have ever done").

Mithoefer and colleagues interpret their findings as evidence that MDMA-assisted psychotherapy produced enduring, clinically meaningful benefits in a cohort of previously treatment-resistant PTSD patients, with many participants maintaining symptom reductions for an average of nearly 3½ years after treatment. The authors present two perspectives on sustained improvement: if the three CAPS non-completers are assumed to have retained improvements based on their questionnaire responses, up to 89% (17/19) showed long-term improvement; using a conservative intent-to-treat assumption that treats the three non-completers as relapsed, 74% (14/19) still demonstrated meaningful, sustained reductions in CAPS scores. The investigators note that the relapse rate observed (11% among CAPS completers; 26% in the intent-to-treat analysis) is within the range reported by other long-term PTSD follow-up studies. They highlight two reassuring safety signals from the LTFU questionnaire: no participant reported harm from study participation and no new substance dependence emerged following MDMA-assisted therapy. Reports of preserved or improved cognitive function at LTFU were consistent with neuropsychological testing carried out in the original trial, addressing concerns raised by some recreational-use and animal studies about potential MDMA-related neurocognitive risks. Several important limitations are acknowledged. Follow-up was incomplete for three MDMA-treated participants, the LTFU lacked a control group because most participants eventually received active treatment, and the interval between treatment and follow-up varied widely because the LTFU was added after study initiation. Confounding by subsequent treatments is also emphasised: at LTFU, 8 of 19 participants were still in psychotherapy and 12 of 19 were taking psychiatric medications, so it is not possible to attribute durability of benefit solely to the original MDMA-assisted sessions. The small sample size limited statistical comparisons between those who continued treatment and those who did not. Blinding was also compromised in the original trial, which complicates separation of drug-specific effects from non-specific therapeutic factors in drug-assisted psychotherapy. To address these issues, the authors describe ongoing and planned work: a randomised, three-arm study in military veterans using low, medium and full doses of MDMA to improve blinding and estimate dose–response; inclusion of measures such as the Posttraumatic Growth Inventory and instruments for mystical or transformative experiences; a proof-of-principle trial testing whether an additional open-label MDMA session can restore prior gains in relapsed participants; and replication studies by other research teams using the published treatment manual.

The investigators conclude that, in this small, treatment-resistant cohort, a course of MDMA-assisted psychotherapy was associated with sustained PTSD symptom reduction for many participants, without evidence of subsequent drug abuse or neurocognitive decline. They regard the results as indicating a favourable long-term risk–benefit profile for a limited number of controlled, clinical MDMA sessions delivered with psychotherapeutic support, and they advocate further controlled and independently replicated research to validate these findings and clarify clinical applicability.