Definium Therapeutics announced positive topline results from Voyage, its first Phase 3 trial of DT120 ODT in generalized anxiety disorder, on August 12. The data itself is strong: a placebo-adjusted 5.4-point reduction in Hamilton Anxiety Rating Scale score at week 12, a standardized effect size of 0.81, and symptom separation from placebo visible by day two. GAD affects an estimated 26 million American adults and has not seen a new FDA-approved drug since 2007. On its own, that would be a solid, single-trial pharma story. What actually makes Voyage worth this desk’s attention is that it is not a single trial. It is Definium’s third consecutive positive readout for the same compound, and it lands alongside a design detail in the company’s next GAD trial that this desk has been waiting to see a sponsor actually build.

Three readouts, not one

Voyage follows Emerge, Definium’s positive Phase 3 result in major depressive disorder announced in June, which itself followed a positive Phase 2b readout in GAD. That is three separate positive results, across two indications, without an interruption. This desk has covered enough psychedelic drug-development programs this year to know how unusual that consistency actually is. Compass’s own Phase 2b program in treatment-resistant depression saw response rates decline from roughly 37 percent at the start of dosing to 20 percent by week 12, a pattern this desk flagged directly at the time. Resilient’s MDMA program, the field’s most clinically mature by trial history, was rejected outright in 2024 over methodological concerns documented for years before the decision. A sponsor stringing together three positive readouts in a row is not, on its own, proof that DT120 will clear every remaining hurdle, but it is a genuinely different track record than most of this desk’s coverage has had occasion to report.

The design detail worth reading closely

Panorama, Definium’s second pivotal GAD trial, reads out in September, and its design is where this story connects directly to methodology this desk has tracked all year. Panorama randomizes participants 2:1:2 across a 100 microgram dose, a 50 microgram dose, and placebo, and Definium’s own release states the low-dose arm is “intended to confound participants’ ability to accurately assess the dose condition to which they have been randomized.” The company goes further, stating this approach “aligns with FDA guidance on the use of complementary designs” across its DT120 program. That is a direct, specific reference to exactly the active-comparator strategy this desk examined in detail when FDA finalized its psychedelic clinical trial guidance in July, guidance that recommended low-dose active comparators specifically to manage the functional unblinding problem that has undermined confidence in results across this drug class. Definium is not gesturing at this concern. It has built a trial arm specifically to test whether its own Phase 2b results held up once dose-guessing became harder, and it is publicly framing that design choice as informed by the agency’s own stated preferences.

What Voyage’s structure still leaves open

Voyage’s double-blind period runs 12 weeks, after which participants enter a 40-week open-label extension where they may receive up to four additional doses based on symptom severity. That structure, a fixed double-blind endpoint followed by an open-label extension with symptom-triggered retreatment, is close in shape to the design this desk scrutinized in Compass’s COMP006 durability data, where a retreatment framework built after the fact, rather than prespecified and blinded, made the six-month durability claim harder to read cleanly. Nothing in Voyage’s topline release indicates Definium has solved that specific problem differently than Compass did. The same standard applied there deserves to be applied here: durability data drawn from an open-label extension with clinician-driven retreatment is a real signal worth tracking, not yet the same evidentiary weight as a prespecified, blinded retreatment design would carry.

Updated August 12, 2026: a contrasting case landed the same week

Resilient Pharmaceuticals, the company formerly known as Lykos and before that MAPS Public Benefit Corporation, has resubmitted its MDMA-assisted therapy NDA for PTSD, according to an August 9 report from Psychedelic Alpha independently confirmed by this desk, and separately addressed in an August 10 statement from MAPS itself. The resubmission proceeds without a new Phase 3 trial, built on the original completed data underlying the program’s 2024 rejection, with proponents arguing the trial’s core efficacy and adverse event findings remained accurate despite the functional unblinding the FDA’s complete response letter specifically flagged as a deficiency. MAPS, which funded and initiated the original clinical program, has no current operational role in Resilient’s development strategy.

The contrast with Panorama is direct and worth naming plainly. Definium built a new trial arm specifically to generate fresh evidence on whether its GAD result survives once functional unblinding becomes harder to achieve. Resilient is resubmitting on its existing data and arguing the same underlying concern, in a different indication and compound, did not meaningfully bias its result in the first place. Both sponsors are answering the same open question this desk has tracked across the psychedelic drug class all year. They are answering it in opposite ways, one with new data built to test the concern directly, the other with an argument that the concern doesn’t change the conclusion already reached. Which approach the FDA finds more persuasive, for either program, remains to be seen.

Reading the effect size in context

A Cohen’s d of 0.81 is a large effect by the standards psychiatric drug trials typically report, and the day-two separation from placebo is fast even by this drug class’s own rapid-onset reputation. Neither figure has been independently replicated outside Definium’s own trial program, and neither addresses the open question this desk has raised repeatedly across psychedelic drug trials generally: how much of a reported effect this large survives functional unblinding, when patients can often correctly guess whether they received an active psychedelic dose. Panorama’s low-dose arm is specifically built to test that question for DT120’s GAD data. Voyage’s own results do not yet answer it.

The caveats

This is topline data from a single trial, released by the company itself, ahead of full data disclosure, peer review, or FDA review of any kind. No suicidality signal was reported, a genuinely reassuring detail, but Voyage enrolled 214 participants across roughly 35 sites, a real but still modest sample for a condition this common. Panorama’s results, expected in September, will be the more informative test of whether DT120’s GAD effect holds once the low-dose comparator design is actually run, and this desk will read that data with the same scrutiny applied here.

The frame

Most of this desk’s psychedelic drug-development coverage this year has been about reading past an impressive topline number to find the durability question, the unblinding risk, or the safety signal sitting underneath it. Voyage earns a more straightforward read than most: a large, fast-onset effect in a genuinely underserved indication, the third consecutive positive result for the same compound, and a companion trial explicitly designed to stress-test the exact confound this desk has spent the year tracking across the entire field. That does not make DT120 a settled story. It makes Definium one of the few sponsors in this space whose next data readout, Panorama in September, is positioned to actually answer the question its own prior results raised, rather than leave it for someone else to ask.