MDMA-assisted psychotherapy for post-traumatic stress disorder: The devil is in the detail

Research progress: MDMA-assisted psychotherapy for PTSD

To date, there are six published randomized controlled trials (RCTs) of MDMA-AP for PTSD. The first five are small Phase 2 RCTs. Of these, four were completed (Mithoefer et al., 2018, 2011; Oehen et al., 2013; Ot’alora et al., 2018), whereas the fifth was shut down after enrolment commenced, reportedly due to political pressures (Bouso et al., 2008). A pooled analysis of all Phase 2 RCT data – including some unpublished data – has been published (Mithoefer et al., 2019), as have several secondary analyses of Phase 2 data (Feduccia et al., 2019, 2021; Gorman et al., 2020; Marseille et al., 2020; Ponte et al., 2021; Wagner et al., 2017) and two uncontrolled studies of long-term follow-up outcomes (Jerome et al., 2020; Mithoefer et al., 2013). Two small open-label trials were also recently published (Jardim et al., 2021; Monson et al., 2020). Here, we focus on the primary findings from existing RCTs in this field.

Several themes can be extracted from the completed Phase 2 RCTs. First, individual studies reported positive results, with participants in the MDMA groups achieving statistically larger reductions from baseline in symptom scores (using the Clinician-Administered PTSD Scale [CAPS]) relative to placebo (Mithoefer et al., 2011, 2018) and/or descriptively higher clinical response rates (defined by reduction from baseline in CAPS scores) (Mithoefer et al., 2011, 2018; Oehen et al., 2013; Ot’alora et al., 2018). One reported significant effects only in a per-protocol analysis, excluding a placebo-responding participant and three in the high-dose condition who disclosed exclusionary psychiatric conditions during treatment (Ot’alora et al., 2018). Second, where treatment effect sizes were reported, they ranged from medium to very large (Cohen’s d from 0.37 to 2.8; Mithoefer et al., 2011, 2018; Ot’alora et al., 2018). Third, each study reported an acceptable safety profile of MDMA, when administered on limited occasions in a controlled medical context. This is consistent with experience in human laboratory studies (e.g. Bedi et al., 2009, 2010; Vizeli and Liechti, 2017). It should be noted, however, that robust follow-up data supporting long-term safety are not yet available: the two follow-up studies based on Phase 2 data did not include a placebo condition (Jerome et al., 2020; Mithoefer et al., 2013).

Methodologically, Phase 2 RCTs employed a combined pharmacotherapy-psychotherapy model, with MDMA hypothesized to catalyse the effects of psychotherapy alone. Between two and three non-drug preparatory therapy sessions were conducted prior to two to three whole day MDMA- (or placebo-)assisted sessions, each of which was followed by a therapy session the next morning and two to three weekly integration sessions before the next medication-assisted session. A combined male and female therapist team was used in each study.

All completed Phase 2 studies employed 125 mg of MDMA in the active condition, with most participants having access to an optional supplemental dose (62.5 mg) available 1.5–2.5 hours after initial dosing to extend the putative therapeutic window. Two studies also included a lower active dose group (100 mg + 50 mg supplement and 75 mg + 37.5 mg, respectively) in addition to the 125 mg group to examine the dose-dependency (Mithoefer et al., 2018; Ot’alora et al., 2018). An inactive placebo was employed in the first completed Phase 2 study (Mithoefer et al., 2011) whereas a low MDMA dose (25 mg + 12.5 mg, 40 mg + 20 mg or 30 mg + 15 mg, respectively) was subsequently used to improve blinding (Mithoefer et al., 2018; Oehen et al., 2013; Ot’alora et al., 2018).

Based on results from Phase 2 trials, the US Food and Drug Administration (FDA) granted MDMA-AP Breakthrough Therapy Designation (Mithoefer et al., 2019) – designed to expedite approval for treatments of life-threatening or serious conditions, where preliminary data suggest that the medication may be more efficacious than existing options. The FDA pre-approved the design of two planned Phase 3 trials (Mithoefer et al., 2019), potentially further expediting approval in the case of positive Phase 3 results.

Last year saw the publication of the first Phase 3 RCT of MDMA-AP for PTSD (Mitchell et al., 2021), to our knowledge the first Phase 3 trial of a psychedelic medicine approach to date (secondary analyses of this dataset were also subsequently published; Brewerton et al., 2022; Nicholas et al., 2022). This study of 90 participants with severe PTSD used a similar design to the Phase 2 studies, with three preparatory sessions, three medication-assisted sessions and three integration sessions in the 4 weeks between each medication-assisted session. The starting MDMA dose was 80 mg followed by a supplemental 40 mg. The target dose for Sessions 2 and 3 was 120 mg plus 60 mg, with the possibility of withholding either the supplemental dose or dose increases if requested or clinically indicated (this rarely occurred). Results supported the efficacy of MDMA-AP, with participants in the MDMA condition showing a greater reduction relative to placebo in both CAPS scores (the primary endpoint) and functional impairment (secondary endpoint). The treatment effect size for the main outcome was large (d = 0.91), with a medium effect on functioning (d = 0.43). MDMA-AP was well-tolerated and, according to the report, did not produce unacceptable adverse effects or noteworthy safety concerns. The second Phase 3 study of MDMA-AP for this indication is ongoing (NCT04077437).

Limitations and methodological concerns

Results from the trials described above are certainly promising and have attracted exhaustive media coverage, including a feature on the front page of The New York Times (Nuwer, 2021). Below, we discuss some of the methodological challenges in relation to clinical trials of MDMA-AP and their potential combined impact on findings.

Key among these challenges is ensuring adequate blinding, given the robust psychoactive and physiological effects of MDMA (Burke and Blumberger, 2021). Most Phase 2 MDMA-AP studies employed low-dose MDMA as an active placebo, which does appear to improve blinding, although the accuracy of condition guesses by participants and therapists remained well above chance (Oehen et al., 2013; Ot’alora et al., 2018). This approach complicates interpretation, investigating dose-dependency rather than comparison with a pharmacologically distinct control condition. Low MDMA doses may also produce anxiogenic, potentially countertherapeutic effects, possibly artificially boosting the efficacy signal for higher doses (Mithoefer et al., 2019). For this reason, the Phase 3 trials used inactive placebo. In the initial Phase 3 study, this resulted in an estimated 96% of MDMA and 84% of placebo participants accurately guessing their condition (Mitchell et al., 2021). Use of a blinded, independent and centralized rater pool was employed to ameliorate this issue in the Phase 3 trials; however, most participants were readily able to guess their condition based on their own subjective experiences.

Blinding failures raise the prospect of expectancy and demand effects biasing outcomes. A substantial literature shows that patient expectations about the likelihood of treatment success affect responding to psychiatric interventions, including pharmacotherapies for PTSD (Graham et al., 2018). For instance, response rates to antidepressant medications in comparative efficacy trials (where only active medications are available) are reliably higher than in placebo-controlled trials for major depression in adults. This appears to be due to expectancies: in comparative efficacy trials, participants know they will receive an active medication because of the design; they therefore have greater expectations of treatment success, ultimately driving stronger medication effects than those in studies with a placebo arm (Rutherford et al., 2017). Where psychedelic medicine studies fail to adequately blind conditions – as has largely been the case for MDMA-AP – the obvious effects of the medication and their absence under placebo likely inflate the observed treatment effect sizes (see Muthukumaraswamy et al., 2021).

Several factors specific to psychedelic medicine may compound these issues. First, legal and political obstacles have meant that research to date has been driven by an enormous amount of work over decades by passionate advocates for psychedelic medicine (see Mithoefer et al., 2019). This work would not have occurred absent a sincere, powerful, a priori conviction on behalf of advocates that these approaches work (even before the research was done). This raises the possibility of allegiance bias, whereby outcomes of psychotherapy research studies are impacted by the belief systems of the researchers (e.g. Munder et al., 2013). Thus, although their conviction has clearly motivated advocates to move the field forward, it undermines the sceptical, dispassionate approach required in clinical research, with unknown effects on trial outcomes.

Second, funding for most research to date has relied on philanthropy (Muthukumaraswamy et al., 2022). To generate public support, organizations have used robust public messaging around the likely impact of these approaches – messaging that may go well beyond the actual evidence. An example is public-facing material by Mind Medicine Australia (MMA), an Australian advocacy group for psychedelic medicine. A promotional video entitled ‘Everyone Should Have Access to Psychedelic-Assisted Psychotherapy’ on the MMA homepage states that ‘With proper clinical support, psychedelic treatments are safe, and frequently lead to remission after only a short program, and even where current treatments have failed’ (www.youtube.com/watch?v=VlllT_BpaCs; starts at 1 minute 25 seconds). This messaging is premature given the limitations of existing evidence and likely contributes to inflated expectations about psychedelic medicine in research participants and the broader public. Moreover, messaging like this is accessible to the public in a way that a more nuanced discussion may not be.

Third, the ‘colourful’ history of psychedelic medicine, routed in the countercultural movements of the 1960s and 1970s, is undoubtedly interesting, rendering psychedelic medicine a widely popular media topic. Regular media reports have focused on psychedelic therapy, often centred around powerful personal anecdotes (Greenway et al., 2020; Noakes, 2019). These narratives, while compelling, do not offer high-quality evidence for this treatment approach. Such media reports have fed expectations that psychedelic therapy may offer rapid and dramatic relief for patients suffering from mental illness.

Fourth, in common with many psychiatric trials, convenience samples have been used in MDMA-AP studies, with participants volunteering directly for enrolment. It is unclear how representative these participants are of individuals seeking psychiatric care through more routine channels. Of note, in the Phase 3 MDMA-AP trial, 32% of participants reported prior MDMA use, which is markedly higher than general population rates in the United States (SAMHSA, 2020), where most participants were recruited. Given the nature of the treatment, which involves 8-hour therapy sessions under the influence of MDMA, it would be unsurprising if this approach is particularly appealing to those with prior positive MDMA experiences. The sense of being part of a ‘movement’ with potentially profound impacts – such as FDA approval of a new, paradigm-shifting treatment – could also produce social desirability effects on participant responding.

Issues surrounding selective, non-representative samples and expectancies affect studies in psychiatry more broadly. However, the combination of blinding difficulties, inflated expectancies, selection bias and demand effects arising from media hype, robust advocacy and simplistic messaging around MDMA-AP and other psychedelic psychotherapies has the potential to powerfully impact outcomes. This is highlighted by a recent media quote from a participant in the Phase 3 MDMA-AP trial:

Everything that I read, everything that I heard about, was how this stuff will fix you . . . I wanted to be fixed, and I felt great responsibility. At the time I felt that if I didn’t get better then the FDA was not going to approve it. . .and millions of people were not going to have access to it. . . It was more of reporting how things were getting better, because that’s what I really wanted, and part of me really believed that too. (Dickinson and Mugianis, 2021)

A final methodological issue concerns the psychotherapeutic modality employed in MDMA-AP studies to date. Of note, the FDA and the Australian TGA do not have regulatory oversight of psychotherapeutic approaches, which lay outside of the remit (and expertise) of both agencies. All existing MDMA-AP studies have employed a therapy developed by the Multidisciplinary Association for Psychedelic Studies (MAPS, the advocacy organization that has sponsored all trials discussed above; Mithoefer, 2017). This approach is manualised (Mithoefer, 2017), and an adherence monitoring process has been defined (MAPS, 2021). The therapy itself takes a non-directive approach, with the therapist viewed as an ‘non-invasive empathic witness’ to facilitate the patient’s ‘innate capacity to heal from the wounds of trauma’ (p. 8). During medication-assisted sessions, participants engage in both quiet introspection and therapeutic discussion (Mithoefer et al., 2011). Music is used to evoke deep emotions and ‘nurturing touch’ may also be used, reportedly with explicit participant consent (Oehen et al., 2013). The manual is designed to allow therapists – who may have diverse professional and training backgrounds – to employ a range of techniques, including elements from, for example, cognitive processing therapy, voice dialogue, psychosynthesis, Hakomi, sensorimotor therapy, holotropic breathwork, Jungian psychology, Buddhist psychology and virtual reality (p.12). While the manual describes the approach as ‘standardized’, the possible inclusion of elements from this wide range of therapeutic modalities limits the extent to which standardization, and hence replicability, is possible. Combined with unblinding of therapists, it is possible that certain techniques or approaches may be more likely to be used in the active than placebo conditions while still meeting the broad adherence criteria.

Future directions

Above, we outline key issues that we believe may be impacting the outcomes of existing MDMA-AP research and inflating observed treatment effect sizes. This gives rise to several recommendations for the future.

A key step in future studies should be use of active controls as well as inert placebo to better disambiguate expectancy-driven effects from true therapeutic effects. It is unclear why low MDMA was the only active control in Phase 2 MDMA-AP studies and inactive placebo is being used in Phase 3 studies, when alternative active controls have been used in other psychedelic medicine studies (Grunebaum et al., 2018; Ross et al., 2016), and to blind MDMA in laboratory studies (Bedi et al., 2010; Kirkpatrick et al., 2012). Active control conditions such as psychostimulants (e.g. dexamphetamine) or serotonergic medications (e.g. trazodone) that have overlapping acute psychoactive effects to MDMA (Johanson et al., 2006) without long-term therapeutic benefits for PTSD in this context (i.e. acute doses combined with psychotherapy) would be a more rigorous approach. Formal blinding checks and expectancy measures should also be used to better assess the impact of blinding failures and expectancies on outcomes.

While media-driven hype cannot be controlled, individual research groups can take steps to minimize effects on study outcomes. Researchers in this field should be mindful of their own interactions with the media and their public messaging to avoid further contributing to overblown expectations.

Recruitment direct from existing treatment services (e.g. in the United States, Veteran’s Affairs services; or in Australia, services such as Phoenix Australia’s Traumatic Stress Clinic), rather than based on advertising (particularly materials naming the study medication), may limit selective recruitment of people with inflated expectancies about these treatments. In our own studies where selection bias and expectancies are a concern, we employ separate recruitment/screening and study protocols, such that participants are recruited for clinical trials of novel treatments for their condition (broadly). They do not receive information about the specific study intervention until they are found to be eligible during screening. This approach reduces selection pressures in recruitment.

More research is urgently needed as to the specific features of psychotherapy required to maximize clinical benefit and reduce risk in the context of MDMA-AP. More structured modalities – such as cognitive-behavioural therapies (CBT) – that are more clearly operationalized behaviourally, with a far more extensive evidence base than the MAPS approach, should be adapted for this purpose. First-line treatment for PTSD in adults comprises trauma-focused CBTs, including cognitive processing therapy, cognitive therapy for PTSD, narrative exposure therapy, and prolonged exposure (National Institute for Health and Care Excellence, 2018). Given that increasing the tolerability of exposure to traumatic memories is a hypothesized mechanism for MDMA’s effects in the treatment of PTSD, existing trauma-focused CBTs could be modified for MDMA-AP. For instance, exposure and processing of trauma memories could be concentrated in MDMA sessions, with psychoeducation, safety planning and affective regulation skills development in preparation sessions and restructuring trauma-related cognitions occurring in integration sessions.

Other features of the psychotherapy employed, such as the number of preparatory and integration sessions and the number of therapists used, should be based on evidence rather than historical factors alone. The possibility that therapists can use techniques from multiple different treatment modalities while still maintaining fidelity to MAPS’ treatment almost certainly undermines standardization of treatment. While a detailed adherence rating process has been established (MAPS, 2021), to our knowledge the proportion of Phase 2 and 3 MDMA-AP sessions that were reviewed using this procedure was not included in published reports. Future studies could valuably implement and report upon standardized treatment fidelity assessment procedures.

To limit the influence of advocacy on study outcomes, large, high-quality and independent efficacy trials of these methods are clearly needed: to date, all RCTs of MDMA-AP have been sponsored by a single advocacy organization (MAPS). We also support the need for effectiveness trials and controlled assessment of long-term outcomes. Even after the second Phase 3 study of MDMA-AP for PTSD is completed, and even if – as expected – results are positive, there will remain much that is unknown about MDMA-AP risks/benefits and effectiveness in the real world.

Conclusion

There can be no doubt that improved treatments for most psychiatric conditions are urgently needed. Despite this need, our collective sense of urgency should not pre-empt rigorous and dispassionate scientific evaluation of these treatment approaches, particularly not when most Australians with mental illness do not have access to existing evidence-based care (Productivity Commission, 2020). No one approach is going to be the ‘answer’ to the mental illness epidemic; setting up such unrealistic expectations will likely cause more difficulties in the long run. In this area of psychiatry as in others, science must take precedence over advocacy; powerful testimonials, however moving, are no replacement for a solid evidence base.