Download the Visual
Novel depression drugs meet their endpoint about half the time, mostly at Phase 2
Most trial wins for new depression mechanisms have come at Phase 2 rather than Phase 3. In Sleuth’s June 2026 analysis of 122 novel-mechanism major depressive disorder (MDD) programs that entered the clinic since 2015, 31 of the 61 with a clear Phase 2 or Phase 3 primary readout met their endpoint, and only 8 of those wins came at Phase 3. The one class with a clean record, serotonergic psychedelics, won almost entirely at Phase 2.
For anyone underwriting an MDD asset, a Phase 2 win on a new mechanism is not proof that the mechanism works. It is proof that the drug cleared the easier bar.
Most mechanism classes have more misses than wins
NMDA and glutamate drugs, the mechanism pharma keeps paying up for, met their primary endpoint in 7 of 16 readouts in Sleuth’s June 2026 analysis. The catch-all group of other mechanisms, the most tested of all, met in 8 of 21. Anti-inflammatory drugs met in 2 of 5, and the single orexin readout missed. GABAergic and neurosteroid drugs split 2 and 2, but both of their Phase 3 readouts missed.
Kappa-opioid antagonists went 0 for 2, and both misses were at Phase 3: Neumora’s navacaprant and Johnson & Johnson’s aticaprant. J&J stopped aticaprant’s Phase 3 VENTURA program for insufficient efficacy in March 2025. On June 15, 2026, Neumora discontinued navacaprant after its KOASTAL-2 and KOASTAL-3 Phase 3 trials failed.
Five of the seven classes have more misses than wins, so a mechanism label alone does little to de-risk a depression program. The same gap between Phase 2 signals and Phase 3 proof runs through Sleuth’s AAIC 2026 Alzheimer’s landscape.
Psychedelics’ clean record rests mostly on Phase 2
Serotonergic psychedelics met their primary endpoint in all 12 of their readouts in the June 2026 analysis, but 10 were Phase 2 and only 2 were Phase 3. Across the whole dataset, almost every win is a short-term drop on a depression scale, usually at Phase 2, often after a single dose and often in trials where patients could tell what they had taken.
On June 22, 2026, the day this analysis was published, Definium Therapeutics reported that a single dose of DT120, an LSD-based tablet, met its primary endpoint in the Phase 3 Emerge trial in MDD, with an 8.1-point placebo-adjusted MADRS improvement at week 6.
The new Phase 3 win fits the same profile as the Phase 2 ones: a single dose and a six-week scale change. Durability and blinding remain the questions to diligence before a clean record is read as a proven mechanism.
Programs and readouts are from Sleuth’s June 2026 analysis of novel-mechanism MDD programs that entered the clinic since 2015; 68 had a Phase 2 or Phase 3 primary readout, mostly on the MADRS depression scale, and the 61 clear readouts exclude seven mixed ones. “Met” means the trial’s primary endpoint. Navacaprant, aticaprant and DT120 developments are dated in the text and were checked in September 2026. This page is a dated snapshot.
Frequently asked questions about novel depression mechanisms
How often do novel-mechanism depression drugs succeed in trials?
In Sleuth’s June 2026 analysis of 122 novel-mechanism MDD programs since 2015, 61 had a clear Phase 2 or Phase 3 primary readout: 31 met the primary endpoint and 30 missed. Only 8 of the 31 wins were at Phase 3.
Which depression mechanisms have failed at Phase 3?
In the June 2026 analysis, kappa-opioid antagonists went 0 for 2 at Phase 3 (navacaprant and aticaprant), and GABAergic or neurosteroid drugs missed both of their Phase 3 readouts. NMDA and glutamate drugs and the other-mechanisms group each had one Phase 3 miss alongside Phase 3 wins.
Do psychedelics work for depression?
Serotonergic psychedelics met their primary endpoint in all 12 readouts in Sleuth’s June 2026 analysis, but 10 were Phase 2, mostly short-term scale changes after one or a few doses. On June 22, 2026, the day the analysis was published, Definium Therapeutics reported that a single dose of DT120, an LSD tablet, met its Phase 3 primary endpoint in MDD.
What does “met endpoint” mean in this analysis?
A readout counts as met when the trial hit its primary endpoint, in most cases a change on the MADRS depression scale. It says nothing about durability or later trials, which is why Sleuth separates Phase 2 wins from Phase 3 wins.
