One Dose, Three Trials:
What the LSD Results Do and Do Not Show
Dr. Thomas Lee Abshier, ND ·
On the Phase 3 results for DT120 (oral LSD) in anxiety and depression, presented at the NEI Fall Congress and reported by MedPage Today.
Where I stand
This past weekend in Colorado Springs, at the Neuroscience Education Institute’s Fall Congress, Definium Therapeutics (formerly MindMed) presented the first full medical-meeting account of three Phase 3 trials of DT120, a tablet form of LSD. In all three, one supervised dose of 100 micrograms beat placebo by a wide margin. Two trials were in generalized anxiety disorder, and one was in major depression. Improvement started within a day or two and remained there three months later.
I take these results seriously. They are not a fluke; they have now been replicated, and the effect is larger than what the daily pills we use now achieve. But I do not read them as proof that a chemical event is the cure for anxiety or depression. What they show, more than anything, is that the mind’s state is not locked in place. A single experience, given in a safe setting, can move a person out of a rut that months of daily medication did not. A naturopathic physician should welcome that finding, because it supports something we have said all along: these conditions are responsive to change. They are not life sentences written in brain chemistry.
I have three reservations: whether the trials were truly blind, what remains unknown about long-term safety, and the gap between a carefully run trial and an ordinary clinic. I set them out below. None of them overturns the result. They limit what we should conclude from it.
What the trials found
All three trials gave each person a single dose, or a placebo, and then followed him for twelve weeks. Independent raters scored symptoms without knowing which treatment the person received. Anxiety was measured on the Hamilton Anxiety scale (HAM-A); depression on the Montgomery–Åsberg scale (MADRS). On both scales, higher scores mean worse symptoms.
| Trial | Condition | People | Drug group change | Placebo change | Difference | Response (drug vs placebo) |
|---|---|---|---|---|---|---|
| Emerge | Major depression | 149 | −13.3 MADRS at week 6 | −5.2 | −8.1 | 35% vs 7% |
| Voyage | Generalized anxiety | 214 | — | — | −5.4 HAM-A at week 12 | 43% vs 16% |
| Panorama | Generalized anxiety | 245 | −9.8 HAM-A at week 12 | −4.7 | −5.1 | 32% vs 14% |
A “response” means the person’s score fell by at least half. Some further details:
- Depression (Emerge). People came in with severe depression, an average MADRS score of about 35. The gap between drug and placebo was largest at week one (14.2 points) and was still 7.3 points at week twelve. About one in four on the drug (24%) reached remission at week six, against 3% on placebo.
- Anxiety (Voyage and Panorama). People came in with an average HAM-A score of about 28, which is moderate to severe. In Panorama, 15% of those on the drug reached remission, against 4% on placebo. The effect sizes (Cohen’s d of 0.81 in Voyage and 0.64 in Panorama) are large by the standards of psychiatric drug trials.
- A low-dose comparison. Panorama added a third group that got 50 micrograms, a dose strong enough to be noticed but expected to do less good. That group improved by 3.6 points over placebo, compared with 5.1 for the full dose.
To put this in proportion: a 5-point drop on a scale where the average person started at 28 is real and worth having, but it is not a cure. Roughly three of every four people on the drug did not reach remission. The placebo groups also improved by about 5 points, which reminds us how much of recovery comes from attention, hope, and the passage of time.
For comparison, a 2024 BMJ review of 218 trials found that SSRI drugs taken alone had only a small effect on depression. By the numbers reported, DT120 does better than that, with one dose instead of a pill every day.
The blinding problem
The strongest objection to every psychedelic trial is that a person who has spent six hours seeing the walls breathe knows he did not get the placebo. Once he knows, his hopes can shape what he reports. The person scoring him may also pick up clues. Researchers call this “functional unblinding.” A 2021 review of psychedelic trials found that most had not even measured whether participants could guess their treatment. An FDA adviser cited the same problem when he voted against approving MDMA therapy for PTSD in 2024, and the agency rejected it.
Definium’s answer has two parts. First, central raters scored symptoms without knowing the treatment or even which visit it was. Second, the 50-microgram group in Panorama was meant to show that noticing the drug is not enough: people in that group could tell they had been given something, yet they improved less. Its earlier Phase 2 trial showed the same dose-response pattern. Some independent psychiatrists also point out that functional unblinding is common with benzodiazepines, stimulants, and even ordinary antidepressants, and that in the LSD dose studies the benefit did not rise in step with how obvious the drug was.
My judgment is that this objection has been partly answered but not closed. The dose-response pattern is the best evidence available, and it points the right way. But the 50-microgram group was small (52 people) and was not designed to be tested statistically, so it suggests more than it proves. Some of the measured benefit is probably expectation. Most of it probably is not.
What the safety record shows, and what it cannot yet show
On the day of dosing, nearly everyone on the drug had side effects. In Panorama, about 95% of the full-dose group reported one, compared with 63% in the placebo group. The most common were visual illusions (68%), nausea (37%), and headache (24%). These mostly ended the same day. The company reports no drug-related serious adverse events, no rise in suicidal thinking, and about the same dropout rate as placebo. A session lasts about six hours, and in Panorama 94% of people met the criteria to go home by hour eight.
That is a reassuring short-term record. What the trials cannot tell us is what happens when a person takes the drug again and again over years, and if it is approved, many will. The benefit fades enough that the trials allow repeat doses: in the depression trial’s open-label phase, up to four more doses over forty weeks. LSD also acts on a serotonin receptor in the heart (5-HT2B), the same receptor behind the heart-valve damage that took the diet drug fen-phen off the market. As far as I know, no case of LSD causing valve disease has ever been documented. The concern has been raised mainly about daily “microdosing,” not about a few doses months apart. Even so, it is the right question to ask of a drug that may be given repeatedly for life, and only long follow-up can answer it.
Why the concerns remain plausible
The reservations deserve more than a footnote, for three reasons.
The money. Definium raised roughly $805 million in a public offering in June 2026 and holds about $1.1 billion in cash and investments. Every number in this essay comes from the company’s own topline announcements. None of the three Phase 3 trials has yet appeared in a peer-reviewed journal. That does not make the numbers false, but it means they come from an interested party, and you should read them that way until independent reviewers have seen the full data.
The trial is not the clinic. In a trial, a person is screened carefully, prepared, watched for six hours by trained staff, and followed closely for months. In ordinary practice, the psychiatrist may have thirty minutes, and nobody is paid to help the patient make sense of what he went through. Rachel Yehuda of Mount Sinai, who studies psychedelics, put the danger plainly: “feeling changed is not the same thing as being healed.” Psychiatric drugs have a long history of doing better in trials than they later do in practice.
The pull of the shortcut. It is very appealing to think one afternoon could undo years of anxiety. Hope of that kind inflates trial results and also shapes how a drug is prescribed once it is approved. That appeal is exactly why the evidence must meet a high standard.
The question under the question
Why does this story draw so much attention? I think what people really want to know is whether there is a way out of suffering that does not depend on years of medication or on willpower they no longer have.
The honest answer is that the trials show the door can open. A brain stuck in fear or despair can be shifted, fast. What happens next depends on what the person does once it opens.
This is where my own field has something to say. Medicine has tended to treat depression and anxiety as chemical faults to be corrected by chemical means, while giving little weight to the evidence that they respond to how a person lives. In the SMILES trial (2017), adults with moderate to severe depression who got seven sessions of dietary counseling, aimed at more whole foods and fewer sweets, refined grains, fried foods, and processed meats, reached remission 32% of the time, against 8% for those given social support. People who changed their diet the most improved the most. The 2024 BMJ review mentioned above found that walking, jogging, yoga, and strength training all had meaningful effects on depression, larger than SSRIs alone.
I hold those studies to the same standard I have applied to Definium. They are not perfectly blind either: you cannot hide from someone that he is changing his diet or going for runs. SMILES was small (67 people), and the BMJ authors rated their own confidence as low. But their effects are of the same general size as DT120’s, without a hallucinogen, without a six-hour supervised session, and with side effects that tend to improve the rest of the body.
So I do not see the LSD result as a rival to natural medicine. I see it as confirmation that the condition is movable, and as a reason to put the foundations in place: diet, movement, sunlight, sleep, honest counsel, and a life with meaning. If a person does receive this drug someday, those are what turn a changed afternoon into a changed life.
What I would tell a patient today
- It is not available yet. DT120 is not approved. Definium plans to meet with the FDA this quarter and to file its application in the first half of 2027. A decision would come after that.
- Street LSD is not this drug. The trials used a measured pharmaceutical dose, screened participants, and kept a medical team in the room throughout. Unknown purity and dose, no screening, and no supervision are a different situation with different risks. Whatever these trials found, they did not test that.
- If it is approved, it may be a real tool for the person who has been stuck despite honest effort. I would want it used with careful preparation, follow-up, and the foundations above, not as a stand-alone event.
- Start with what you already have. Changes in diet and daily movement have evidence behind them, cost little, and harm no one. Most people never give it a serious trial.
Sources and their limits
I could not read the MedPage Today article itself, so this essay draws on the company’s announcements and on news coverage of the same trials. The trial numbers come from Definium Therapeutics press releases (Emerge, June 22, 2026; Voyage, August 12, 2026; Panorama, September 14, 2026; NEI Fall Congress, November 4, 2026), as reported by Psychiatric Times, HCPLive, Fierce Biotech, Healio, and Clinical Trial Vanguard. Commentary on blinding comes from Muthukumaraswamy and colleagues (Expert Review of Clinical Pharmacology, 2021), Rosenbaum (Journal of Clinical Psychiatry, 2024), and Psychiatric Times (October 30, 2026). The heart-valve question draws on Tagen and colleagues (2023) and on Bryan Roth’s work on the 5-HT2B receptor. The lifestyle evidence is Jacka and colleagues, BMC Medicine 2017 (SMILES), and Noetel and colleagues, BMJ 2024.
These sources lean on the company’s own topline record and on trade and medical press that work largely from it. The full trial data, including how often participants correctly guessed their treatment, have not been published. If anything has been left out of the topline summaries, it would most likely be in the blinding data, the people who dropped out, and the long-term follow-up. Those are the first things I will look for when the full papers appear.