Guided, Then Penalized: A Deep Dive into the FDA’s MDMA Ruling and What it Means for Psychedelic Clinical Trials

In August 2024, the FDA rejected the New Drug Application for MDMA-assisted therapy for PTSD, sponsored by Lykos Therapeutics, despite two positive Phase 3 trials (MAPP1 and MAPP2), both pre-dating the FDA’s 2023 draft guidance, and years of collaborative engagement between the sponsor and the agency. The decision surprised many across the psychedelic clinical trials field and raised serious concerns about how the FDA Advisory Committee applied regulatory expectations that evolved after the trials were completed.

Sponsors across the industry are now redesigning their trials because FDA’s regulatory posture remains opaque and unpredictable. This report dissects the case of MDMA: what the trial got right, where the FDA shifted its stance, and how the fallout is shaping the next generation of psychedelic clinical trials. More importantly, it lays out what the FDA must do next to restore sponsor trust and ensure that regulatory rigor does not come at the cost of scientific and therapeutic progress.

Trial Design Review: MDMA’s Protocol Through the FDA Lens

The Phase 3 program submitted by Lykos Therapeutics (formerly MAPS PBC) comprised two randomized, double-blind, placebo-controlled studies—MAPP1 and MAPP2—to evaluate the efficacy and safety of MDMA-assisted therapy for adults with moderate to severe PTSD. MAPP2, published in Nature Medicine in 2023, served as the confirmatory pivotal trial supporting the New Drug Application (NDA), while MAPP1 (2021) provided foundational evidence of efficacy. The design of both trials was pre-approved by the FDA through a Special Protocol Assessment to generate data to support an NDA, adhered to Good Clinical Practice guidelines, and was conducted under FDA, DEA, and IRB oversight.

In both studies, participants received three 8-hour experimental sessions approximately one month apart, taking either MDMA (80–120 mg, with optional supplemental dosing of 40–60 mg 1.5–2.5 hours later) or an inactive placebo, under the supervision of two trained therapists. A 1:1 computer-generated block randomization scheme stratified by site ensured balanced group allocation and minimized bias. To further mitigate bias, the sponsor implemented “firewalls” between efficacy data and both site and sponsor personnel, established oversight from an independent data monitoring committee, and ensured all therapists were trained to reinforce the possibility of a range of effects—including no perceptual changes. Informed consent forms explicitly listed alternatives to study participation, and participants were primarily recruited via referrals from neighboring PTSD clinics and treating physicians. Each treatment session was embedded within a manualized psychotherapeutic protocol that included preparatory and integration sessions, structured consistently across both arms.

Efficacy was measured using the Clinician-Administered PTSD Scale for DSM-5 (CAPS-5), administered at baseline and follow-up intervals by independent, video-based raters who were blinded to treatment assignment. The protocols emphasized that using centrally trained, blinded clinicians to conduct these structured interviews—with inter-rater reliability procedures in place—helped minimize bias more effectively than participant-reported outcomes or therapist-led assessments. This approach was specifically chosen to prevent functional unblinding from influencing primary endpoints, as clinicians were never present during drug administration. Placebo capsules were visually identical and administered in a matched therapeutic context to control for expectancy and environmental cues.

MAPP1 (N=90) and MAPP2 (N=104) both demonstrated statistically significant reductions in PTSD symptoms for MDMA compared to placebo. In MAPP1, the MDMA group showed a mean CAPS-5 reduction of −24.4 versus −13.9 for placebo (p<0.0001). In MAPP2, the reduction was −23.7 versus −14.8 (p<0.001). Across trials after three MDMA or placebo sessions, approximately, 88% in MAPP1 and 87% in MAPP2 of MDMA participants experienced clinically meaningful improvement, and 67% of MAPP1 and 71% of MAPP2 patients no longer met PTSD diagnostic criteria at study end, compared to 32% (in MAPP1) and 48% (in MAPP2) in the placebo group.

Frequently reported adverse events reflected known MDMA effects—muscle tightness, decreased appetite, nausea, sweating, dry mouth, feeling cold, elevated blood pressure and pulse—but no serious drug-related adverse events occurred. Cardiovascular monitoring was embedded in trial protocols, and all therapists were trained in emergency response procedures. Safety follow-up continued for 12 weeks post-treatment to monitor ongoing risks and durability.

Case Study: The FDA Advisory Committee Meeting and the Lykos Rejection

The June 4, 2024 meeting of the FDA’s Psychopharmacologic Drugs Advisory Committee (PDAC) to review Lykos’ MDMA-assisted therapy submission revealed a critical disconnect between regulatory expectations and guidance implementation. Rather than focusing on traditional metrics like safety signals or statistical efficacy, the committee scrutinized elements such as the psychotherapy protocol, functional unblinding, and trial design choices—many of which were either ambiguously addressed or not explicitly required by the FDA’s own 2023 draft guidance.

A number of panelists raised concerns about the inability to parse MDMA’s effects from the psychotherapeutic context, criticizing the lack of a dismantling or factorial design. One panelist remarked, “People felt it should have been structured. It should have been a two by two factor design to understand it better,” while another added, “I tend to be a strong proponent of two by two factorial studies in general… there may be other ways to try to address or tease things out.” Some questioned whether a more directive, highly manualized therapeutic approach might have mitigated perceived risks, referring to Lykos’ model as “spontaneous.”

The therapeutic approach used in the MAPP trials has been described as participant-led, supportive psychotherapy—structured yet non-directive, emphasizing patient agency and empathetic engagement. The framework emphasized present-focused, empathetic support that prioritized patient agency, with therapists trained to provide structure without imposing content. As one panelist noted, “The agency’s points about the variability primarily applied to the integrative psychotherapy sessions that do not include drug,” suggesting the critique centered more on post-dosing consistency than the overall therapeutic model.

Further scrutiny centered on the possibility of functional unblinding. One panelist remarked, “There really should have been a blinding assessment tool included… it’s not just about who guesses what, but whether participants knew based on subjective effects.” Another added, “This omission makes it hard to rule out expectancy bias.”

However, a formal blinding questionnaire was conducted during MAPP2 and included in the published supplement. Before trial execution, the potential for functional unblinding was likely discussed with the FDA, and the protocol incorporated mitigation measures such as blinded, centralized raters and firewalling of efficacy data from site staff. While committee members remained concerned, one panelist acknowledged, “The contribution of the likely expectation bias cannot be discounted, while it also cannot be quantified.”

Perhaps most dramatically, public commentary featured allegations of therapist misconduct, including a Phase 2 participant who described physical and sexual assault during the trial. One commenter stated, “It was years before I even realized what had happened to me was abuse, and no one ever followed up.” A panelist later reflected on this, saying, “This raises fundamental questions about oversight and the sponsor’s duty to monitor and report.”

However, this incident occurred during a Phase 2 study and was reported to FDA as both an ethical violation and an adverse event. The case was included in the advisory committee briefing materials (page 86 of FDA briefing document) and reviewed during regulatory inspections, which resulted in process improvements that were included in the NDA. While the incident was not part of the Phase 3 data package, its resurfacing during the advisory meeting cast a lingering shadow over public perception of Lykos’ training and monitoring protocols, despite evidence that corrective steps had already been taken and disclosed.

Despite these concerns, several experts spoke passionately in favor of the therapy, including veterans and clinical psychiatrists. One expert stated, “Approval for MDMA assisted therapy will go down in history as a pivotal milestone… Alleviating human suffering is a moral responsibility.” Another emphasized, “The existing treatments are failing a large portion of our patients. This offers something profoundly new and potentially life-saving.” These voices countered the critiques by asserting that the trial demonstrated safety and efficacy, emphasizing the urgent need for innovative approaches given the limitations of current PTSD treatments.

Yet the final vote reflected hesitation: even those who acknowledged the data’s promise expressed fear that the current infrastructure could not ensure patient safety at scale. As one panelist noted, “The system isn’t yet ready to handle this kind of treatment safely at scale—even if the science is promising.” This reveals a deeper institutional issue—not merely with MDMA or Lykos, but with the FDA’s preparedness to assess therapies that blend pharmacology and psychotherapy in unprecedented ways.

Policy Breakdown: Where FDA’s Psychedelic Clinical Trials Guidance Misguided

The FDA’s 2023 draft guidance on psychedelic drug development was released after Lykos had completed both MAPP trials under a Special Protocol Assessment. Yet, in the Advisory Committee’s 2024 review, the guidance appeared to serve as an implicit benchmark—despite being issued years after trial design was finalized. This created an implicit mismatch between FDA-endorsed design at the time of trial approval and the expectations applied during regulatory evaluation. This timeline is critical. Sponsors were operating in a vacuum, relying instead on case-by-case feedback from the agency, which proved insufficiently predictive of how the final submission would be judged.

What the guidance offered—after the fact—were flexible, nonbinding recommendations. But those same suggestions became the lens through which advisory committee members retroactively scrutinized Lykos’ design, effectively holding the sponsor to standards that didn’t exist when the studies were conceived or approved. The comparison table below highlights this disconnect.

Note: The FDA’s 2023 draft guidance was published after the completion of both MAPP trials. This comparison is not meant to judge Lykos or the FDA or its committee retroactively, but to illustrate the evolving standards that the FDA Advisory Committee seemed to apply post hoc. It highlights where trial expectations may have changed between the time of design approval and final review.

Excerpt from FDA Draft Guidance (June 2023)	What Lykos Did	Did It Meet Guidance?	FDA Evaluation at Advisory Committee (in context of Guidance)	Did the Committee Contradict or Align with Guidance?
“Sponsors should plan to justify the inclusion of a psychotherapy component and describe any trial design elements intended to reduce potential bias or to quantify the contribution of psychotherapy to the overall treatment effect.“	Lykos included a manualized therapy protocol in both arms, with standardized training and integration sessions, but did not justify or measure the independent contribution of therapy.	🔴 No — therapy was justified broadly but not evaluated as an independent variable	❌ Committee questioned the influence of unmeasured therapy effects on efficacy outcomes	❗️ Guidance lacked specificity; committee treated therapy as a confounder rather than a co-component
“Sponsors should consider the use of video or central raters blinded to treatment allocation and visit number.“	Lykos employed blinded, independent video raters for efficacy assessments via CAPS-5.	🟢 Yes	✅ Committee acknowledged this aligned with guidance expectations for reducing bias	✅ Aligned with guidance
“The use of a blinding questionnaire for both subjects and investigators/raters can be helpful to evaluate the impact of functional unblinding.“	Lykos administered a formal blinding questionnaire in MAPP2, consistent with guidance, and prospectively discussed functional unblinding risks with the FDA.	🟢 Yes — survey was implemented and reviewed	❌ Committee still raised concerns about functional unblinding and expectancy bias	❗️ Committee appeared to dismiss measures that met guidance, treating the issue as unresolved
“Observation by two monitors for the duration of the treatment session A healthcare provider with graduate-level professional training and clinical experience in psychotherapy, licensed to practice independently, serving as the lead monitor.“	All sessions were conducted with two trained therapists, including at least one licensed mental health professional.	🟢 Yes	✅ Acknowledged adequate in-session monitoring, consistent with guidance language	✅ Aligned with guidance
“Because psychedelic drugs have serotonin (5-HT) activity, a thorough evaluation of binding to 5-HT receptor subtypes should be conducted.”	Pharmacology studies submitted included 5-HT2B receptor data and safety assessments.	🟢 Yes	✅ No major objections; considered aligned with preclinical requirements in guidance	✅ Aligned with guidance
“Complementary trial designs… could address different challenges in psychedelic drug development. For example… factorial design.”	Lykos used a two-arm design (MDMA+therapy vs. placebo+therapy). A drug-only arm wasn’t feasible due to ethical and blinding limitations.	🟡 Partial — factorial design not used, but design constraints justified	❌ C ommittee highlighted lack of clarity about which component drove efficacy; suggested a factorial approach should have been used	❗️ Committee treated an illustrative example as expected practice, without regard for practical constraints

Where the Conflicts Emerged

Lykos designed both MAPP trials in close consultation with the FDA and secured a SPA in 2017, which signaled formal agreement on trial methodology and endpoints. These discussions occurred years before the agency issued its 2023 draft guidance on psychedelic drug development. While Lykos could not have followed guidance that did not yet exist, its protocol nonetheless reflected many of the document’s later recommendations — including the use of blinded central raters, firewalled data review, and a formal blinding questionnaire in MAPP2.

The disconnect arose during the 2024 Advisory Committee review, where panelists retroactively applied expectations that evolved after the trials were completed. Moreover, in some cases, the committee treated optional guidance elements as mandatory, or introduced entirely new standards that had not been part of prior FDA communications.

This created a twofold problem: not only was the guidance used retroactively, but it was also interpreted inconsistently, and appeared to go against SPA recommendations, undermining the integrity of FDA’s review process and eroding sponsor trust.

1. Psychotherapy Component: The FDA guidance encourages sponsors to justify the inclusion of psychotherapy and describe any design elements used to reduce bias or evaluate its contribution—but does not require isolating the effect of therapy. Lykos included a standardized, manualized therapy protocol in both arms but did not treat therapy as an independent variable. The committee criticized this as a major limitation, applying expectations that aligned more with the later guidance than with prior FDA feedback.
2. Blinding Validation: While the guidance states that blinding questionnaires “can be helpful,” it does not mandate them. Lykos conducted a formal blinding survey in MAPP2, which was reviewed and published. The FDA had also approved this approach in advance. Nevertheless, the committee questioned the adequacy of the blinding protocol, effectively treating an optional recommendation as a requirement.
3. Trial Design Alternatives: The guidance mentions factorial or dismantling designs as illustrative examples—not as requirements. Lykos used a two-arm design (MDMA+therapy vs. placebo+therapy), with therapy in both arms to maintain blinding and meet ethical standards. The committee’s critique that the design failed to isolate treatment components overlooked the practical constraints of doing so and again appeared to treat an example as an expectation.

While Lykos’ trial design may not have been perfect, it’s critical to recognize that the deeper issue lies with the FDA’s inconsistent application of its own guidance. Notably, Lykos had engaged with the FDA extensively throughout the trial design process, including securing a Special Protocol Assessment agreement and Breakthrough Therapy Designation in 2017, indicating prior agreement on key aspects of the study design. Additionally, the agency’s 2023 draft guidance on psychedelic drug development repeatedly emphasized flexibility, describing most recommendations as nonbinding and explicitly stating that psychotherapy components are acceptable if justified. Nowhere did it mandate a dismantling trial, and while it recommended blinding questionnaires as optional tools, Lykos in fact implemented one in MAPP2—at the FDA’s request and in line with the guidance. Yet, both issues were central to the advisory committee’s critiques, revealing a deeper inconsistency between written guidance, FDA approval during the SPA, and how it was interpreted in review.

This was neither a failure of sponsor compliance nor of FDA policymakers. It was a failure of regulatory coherence and consistency—where committee-level interpretation diverged from previously endorsed FDA pathways. Despite engaging with the FDA, sponsors paid the hefty price of running Phase 3 clinical trials only to be rejected later.

By issuing guidance that uses permissive language (“should consider,” “can be helpful“) but then later enforcing those elements as if they were hard requirements, the FDA created a regulatory bait-and-switch. Sponsors like Lykos relied on FDA feedback during protocol development and hoped that meeting the guidance’s suggestions would suffice. Instead, they were penalized for not exceeding them.

The consequences of this ambiguity are not limited to Lykos. Developers now face a chilling effect: if even a Breakthrough Therapy designation and successful Phase 3 results aren’t enough, what is? The lack of transparency in how guidance translates into regulatory expectations is now the single biggest barrier to progress in psychedelic medicine.

If FDA intends to raise the bar for this class of treatments, it must do so explicitly and prospectively. Anything less risks undermining sponsor trust, wasting millions in clinical investment, and delaying access to potentially transformative therapies for patients in need.

Industry Pivot: How Sponsors Are Redesigning Trials Post-MDMA

Lykos’ failed MDMA submission has triggered a wave of risk-averse trial redesigns across the psychedelic field. But the shift isn’t just about sponsor caution—it’s about FDA inconsistency.

The agency’s 2023 draft guidance framed trial design suggestions as flexible. Yet the advisory committee penalized Lykos for not meeting never formally required standards. This gap between written guidance and regulatory enforcement has made FDA expectations unclear, and now, sponsors are overcorrecting to avoid the same fate.

Here’s how they’re adapting:

Strategic Shift	Example / Industry Response
Isolating Drug Effect from Psychotherapy	Several sponsors, including MindMed and atai Life Sciences are adapting trial designs to minimize or eliminate the psychotherapy component to isolate the pharmacological effects of their compounds. MindMed’s recent LSD trial for anxiety explicitly excluded psychotherapy, while atai’s DMT-based study uses a streamlined, time-limited clinical paradigm that emphasizes drug action over therapeutic interaction.
Blinding Validation and Expectancy Controls	Sponsors are incorporating active placebos (e.g., niacin, low-dose psychedelics) and formal blinding questionnaires to reduce bias and measure participant expectations.
Factorial and Dismantling Designs	Several trials now use factorial or dismantling designs—e.g., drug-only, therapy-only, and placebo-only arms—to isolate treatment effects. For example, Awakn’s four-arm trial (ketamine ± therapy, placebo ± therapy) exemplifies factorial designs used to isolate treatment components.
Standardized Therapy Protocols	Sponsors are under growing pressure to standardize therapy protocols and implement therapist training and fidelity checks to mitigate risks and satisfy FDA concerns.
Preemptive REMS Proposals	Sponsors like Lykos are proposing REMS-style frameworks—including certified sites, trained providers, and in-clinic-only dosing—to address FDA safety expectations.

Restoring Trust: What FDA Must Do Next

The rejection of MDMA-assisted therapy exposed systemic flaws in FDA’s regulatory posture. Although the agency’s draft guidance was issued after Lykos completed its trials, the Advisory Committee retroactively applied its suggestions as if they were binding standards. In some cases, Lykos had already incorporated those very safeguards—such as blinded raters and a formal blinding questionnaire—yet the committee dismissed them or treated them as insufficient. This created a disconnect between historical FDA engagement and final review expectations, effectively penalizing a sponsor for not anticipating future regulatory preferences. If FDA wants to foster innovation while maintaining scientific rigor, it must be consistent. Here’s where to start:

Clarify Trial Design Expectations—Explicitly

The agency’s guidance encourages—but does not require—dismantling or factorial designs. Yet the Advisory Committee treated their absence as a critical flaw. FDA must clarify whether such designs are now expected, and under what circumstances they are optional. If expectations have changed since prior trial approvals, that shift must be communicated proactively and prospectively.

Sponsors can’t plan Phase 3 trials around moving targets. Clear, forward-looking standards reduce waste and improve predictability.

Define Acceptable Limits for Functional Unblinding

Functional blinding was a central concern during the advisory review—even though FDA has previously acknowledged that full blinding is often infeasible in psychedelic trials. But without clear benchmarks, sponsors are left to guess what will satisfy regulators. Whether it’s through expectancy questionnaires, sensitivity analyses, or rater-blinding protocols, the agency must define what it considers acceptable mitigation—especially in an environment where the standards continue to shift. The lack of concrete expectations turns functional blinding into a moving target—leaving even well-prepared sponsors vulnerable to post hoc criticism.

Without clear criteria, valid trial designs risk being undermined by subjective interpretation.

Create a Review Framework for Drug + Therapy Combinations

MDMA’s efficacy was evaluated as if it were a standalone drug, despite its intended use within a structured psychotherapy protocol. FDA must develop a regulatory model for interventions where drug and therapy are co-primary elements—much like fixed-dose combinations in oncology or HIV.

Expecting a pharmacologic-only effect from inherently integrative treatments is both inconsistent and scientifically limiting.

Standardize REMS Expectations for Psychedelics

Sponsors are already proposing REMS-like systems—certified providers, in-clinic dosing, therapist training—to meet anticipated FDA safety concerns. The agency should formalize these expectations in advance. Drawing from models like Spravato, FDA can help ensure that safety strategies are aligned before pivotal trials begin.

This would reduce last-minute objections and enable earlier consensus around risk mitigation.

Will EMA Lead Where FDA Hesitates?

While the FDA’s draft guidance has sparked confusion due to its vague and nonbinding suggestions, and has been inconsistent in its decisions on psychedelic therapies, the European Medicines Agency (EMA) has taken early steps toward shaping a more collaborative and transparent approach. In its multi-stakeholder workshop on psychedelics, the EMA acknowledged many of the same challenges—such as functional unblinding, standardization of psychological support, and the need for adaptive trial designs—but added new dimensions to the conversation. Specifically:

Pragmatic Differentiation Between Psychotherapy and Psychological Support

EMA made a clear distinction: if the non-drug component is purely for patient safety (e.g., sitter model), it can be protocolized. But if it’s meant to be therapeutic, it must be studied as a co-primary treatment—emphasizing the importance of intent and trial labeling 

Factorial Designs

EMA recognized that if psychotherapy is considered an active component (not just a safety net), trials may require a factorial design to properly disentangle treatment effects—mirroring FDA’s emerging stance, but with more direct acknowledgment of feasibility concerns raised by industry.

AI & Standardization Tools

Unlike the FDA, EMA stakeholders discussed leveraging artificial intelligence to ensure consistent therapist behavior and adherence to protocols across study sites—a possible future mitigation strategy for reducing variability in psychotherapy delivery.

Guideline Evolution

EMA noted that its depression trial guidance was recently updated to include psychedelics and may evolve further as new data emerge. Developers were encouraged to use the Scientific Advice process to clarify expectations prospectively.

Summary: A Tipping Point for Psychedelic Regulation?

The FDA’s rejection of MDMA-assisted therapy—despite two positive Phase 3 trials and years of engagement—exposed a critical misalignment in the agency’s regulatory process. Lykos followed an FDA-endorsed trial design under a Special Protocol Assessment. But when the data came up for review, the Advisory Committee evaluated it against expectations that had not been formalized at the time of trial initiation.

This disconnect may not reflect failure on any single group—but it reflects a system-level breakdown in how evolving standards are communicated, applied, and aligned across stakeholders. When sponsors receive FDA approval for trial design, that alignment should carry weight through final review. Otherwise, the regulatory goalpost moves midgame—undermining trust, wasting resources, and delaying access to promising treatments.

To restore clarity and confidence in psychedelic drug development, FDA must ensure that SPAs, guidance, and committee-level expectations speak in one voice.