Opparounds

Efficacy and Safety of Psilocybin in Treatment-Resistant Major Depression: The EPISODE Randomized Clinical Trial

Mertens LJ, et al. · JAMA Psychiatry · Published online March 18, 2026

EPISODE randomized 144 adults with treatment-resistant major depression to a single session of psilocybin 25 mg, psilocybin 5 mg, or nicotinamide as an active control, each with psychological support, and followed them for six weeks. The primary outcome was response — a 50% or greater reduction in HAMD-17 — and on that measure the trial did not succeed: 17.0% responded to 25 mg versus 12.5% to 5 mg and 10.6% to nicotinamide, an adjusted odds ratio of 1.73 (95% CI 0.53 to 6.23), P = .19.

The secondary continuous outcome told a different story. Mean HAMD-17 change from baseline at week 6 favored the 25 mg arm by 4.60 points versus control, P < .001 — a difference that is clinically meaningful even though the responder analysis was not statistically significant. Adverse events occurred in essentially every participant given 25 mg, mostly acute and mild-to-moderate. Suicidal ideation was reported in 4% on dosing day versus 1–2% of comparators, and there were two serious adverse events attributed to the 25 mg dose, one of them hallucinogen persisting perception disorder.

Read the paper →  doi:10.1001/jamapsychiatry.2026.0132

Phase 3 Trial of Semaglutide in Metabolic Dysfunction–Associated Steatohepatitis (ESSENCE)

Newsome PN, Sanyal AJ, et al. · New England Journal of Medicine · 2025

ESSENCE randomized adults with biopsy-confirmed MASH and stage 2–3 fibrosis to once-weekly subcutaneous semaglutide 2.4 mg or placebo, with paired liver biopsies at 72 weeks. Resolution of steatohepatitis without worsening fibrosis occurred in 62.9% of 534 patients on semaglutide versus 34.3% of 266 on placebo.

The second primary endpoint, improvement in fibrosis without worsening steatohepatitis, was met in 36.8% versus 22.4% — an estimated difference of 14.4 percentage points (95% CI 7.5 to 21.3, P < .001). Safety was consistent with prior semaglutide trials, dominated by gastrointestinal effects. This is histologic, biopsy-adjudicated improvement rather than a surrogate biomarker, which is what makes it a landmark rather than another weight-loss readout.

Read the paper →  doi:10.1056/NEJMoa2413258

When a clozapine patient quits smoking, the danger is the cigarettes stopping — not the nicotine

Tobacco smoke induces CYP1A2, and clozapine is a CYP1A2 substrate. The inducing agents are the polycyclic aromatic hydrocarbons in the smoke, not the nicotine — which is the clinically decisive detail, because it means nicotine replacement, varenicline, and vaping do not substitute for the induction. A patient who is admitted to a smoke-free unit, or who succeeds at quitting, loses the induction over roughly a week and their clozapine level climbs on an unchanged dose.

The rise is not subtle: plasma concentrations commonly increase on the order of 50 to 100%, which is enough to carry someone from a comfortable therapeutic level into sedation, sialorrhea, myoclonus, or frank seizure. The practical move is to anticipate rather than react — reduce the dose by roughly a quarter to a third over the first week of abstinence and check a level, rather than waiting for toxicity to declare itself. The mirror-image error matters too: a patient who resumes smoking re-induces the enzyme and can drift back to subtherapeutic levels and relapse, which is easily misread as nonadherence. Caffeine inhibits the same enzyme, so a simultaneous jump in coffee intake — common in early abstinence — pushes levels the same direction.

Projective identification, and what containment actually requires of you

Projective identification is not a synonym for projection. In projection the patient attributes something of their own to you and you are largely a screen. In projective identification the patient unconsciously evokes the state in you — you do not merely get accused of contempt, you find yourself actually feeling it. The diagnostic sign is that the feeling is ego-alien: stronger, flatter, or more out of character than your usual countertransference, and often it arrives before you can account for it.

Containment, in Bion's sense, is not stoicism and it is not disclosure. It is holding the evoked state long enough to think about it rather than act it out or hand it back — and then speaking to the relationship it describes, not to your own experience of it. The technical failure modes are symmetrical: acting it out (becoming curt, running over, forgetting the appointment) confirms the patient's model of relationships, while announcing it ("you're making me feel contemptuous") relocates the problem into the patient as an accusation.