New Phase 3 Numbers: 71% of the MDMA Group No Longer Met PTSD Criteria

The most recent Phase 3 figures for MDMA-assisted therapy in PTSD were summarised by UCSF Psychiatry and Behavioral Sciences in October 2026, and the headline is strong: by the end of the 18-week trial period, 71.2% of people in the MDMA-assisted therapy group no longer met the diagnostic criteria for PTSD, against 47.6% of those who received therapy plus placebo. The report states that the medicine produced no serious adverse side effects.

Those two numbers have to be read together, because the second one is the reason the first is contested.

What the control group did

Nearly half the placebo group — 47.6% — stopped meeting criteria for PTSD. That is not a small effect. It is a large, clinically meaningful improvement in people who received the identical psychotherapy without the drug, and it is the reason the FDA's advisory committee found the efficacy case close. When the active arm and the control arm differ by roughly 24 percentage points, the question is not whether MDMA does something; it is whether the trial design can prove the drug rather than the therapy did it.

This is the same structure as the earlier confirmatory trial published in Nature Medicine in 2023, where PTSD symptom severity (CAPS-5) fell by a least-squares mean of 23.7 points with MDMA-assisted therapy versus 14.8 points with placebo plus therapy, and where a majority of the MDMA group — 67% — no longer met diagnostic criteria at week 18 against 32% of controls.

Why the numbers moved

The shift from 67% versus 32% to 71.2% versus 47.6% is worth noticing: the MDMA arm improved slightly, and the control arm improved substantially. A control group that gets better on its own is usually a sign of better-conducted therapy, better-selected sites, or participants who expected more than the trial assumed. Each of those explanations cuts against a simple "the drug works" reading, and each is a legitimate finding rather than a criticism of the patients involved.

What the safety report says

"No serious adverse side effects" is the claim being made, and it is consistent with the published trial record: in the Nature Medicine report, cardiac events such as palpitations and tachycardia occurred but were graded mild, in a screened population under continuous monitoring. The important qualifier is the one the safety literature keeps repeating — this record was produced in people who passed cardiovascular screening, with vital signs tracked throughout, in a two-therapist room.

What is still missing

• Blinding. Participants and therapists can usually tell who received the drug, which inflates unblinded outcomes.
• Active comparator. There is no trial comparing MDMA-assisted therapy against a proven trauma therapy such as EMDR or prolonged exposure.
• Durability. Follow-up beyond 18 weeks is thinner than the acute result, and relapse is the clinical question that matters.
• Generalisability. Trial participants were screened for cardiovascular and psychiatric comorbidity, so the result describes a filtered population.

Where this leaves the application

The FDA rejected the application in August 2024 and asked for a new randomised, double-blind Phase 3 trial plus an independent audit of study records. A resubmission without that new trial was reported in August 2026. These newer figures are the evidence the sponsor is leaning on. Whether an agency that asked for a harder trial accepts a softer one is now the question — and the answer will determine whether patients in the US see this treatment at all.

For readers, the useful takeaway is narrower than the headline: in a screened population, with therapy in both arms, adding MDMA produced more people who no longer qualified as having PTSD. That is a real result, and it is not yet a settled one.