The standard description of ketamine as an antidepressant is that it works by blocking the NMDA glutamate receptor, a mechanism unrelated to the opioid system and therefore free of opioids’ dependence problem. The June issue of the American Journal of Psychiatry, volume 183, number 6, carries a cluster of papers that makes that description harder to defend. Taken together with a decade of prior work, they strengthen the case that ketamine’s antidepressant and antisuicidal effects engage the mu-opioid receptor, and that has consequences for how the drug and its approved cousin are scheduled, marketed, and succeeded.
What the new papers add
Two studies anchor the cluster. The first is a randomized, double-blind, placebo-controlled trial of low-dose sublingual buprenorphine, a partial mu-opioid receptor agonist, given as a follow-on to intravenous ketamine to prolong its effect on suicidal ideation in major depression. The premise stated by the authors is direct: ketamine reduces suicidal ideation quickly but transiently, and one way to extend the benefit is to lean on the mu-opioid receptor pathway that contributes to the antidepressant and antisuicidal response in the first place. The senior author is Alan Schatzberg of Stanford, who has been central to this line of inquiry. The second study is a randomized crossover trial that used naltrexone, an opioid receptor antagonist, to test whether blocking the opioid system changes ketamine’s acute effects on regional cerebral blood flow. It found ketamine increased blood flow in mood-related regions including parts of the anterior cingulate cortex, thalamus, insula, and hippocampus, and examined how opioid blockade altered that signature.
Around those sits a review arguing that ketamine’s pharmacology should be formally redefined to include opioid receptors and the endogenous opioid system, rather than described as NMDA antagonism alone, and proposing that the antidepressant effect arises from a synergistic interaction between the glutamate and opioid systems.
This is not a new claim, and it is not settled
The thesis traces to a 2018 trial, also in this journal, in which Schatzberg and colleagues found that naltrexone blocked the antidepressant effect of ketamine so thoroughly that the study was halted at interim analysis. It was the first time a receptor site had been shown to be necessary for any antidepressant’s mechanism of action in humans, and Schatzberg used the result to caution against widespread and repeated use of ketamine until the mechanism and the risks of tolerance and dependence were better understood. That warning has largely gone unheeded as ketamine clinics proliferated.
The evidence since has been contested rather than conclusive. Critics have argued that naltrexone may carry its own mood effects, that the blockade findings have design limits, and that the opioid effect may not be specific to ketamine. A systematic review published in January 2026 examined 16 studies, including 12 clinical studies covering roughly 790 patients, and concluded that the opioid system’s involvement in ketamine and esketamine’s antidepressant efficacy remains unclear. The honest reading is that the June cluster adds converging clinical and neuroimaging weight to the opioid hypothesis without closing the question. What it does is shift the balance of evidence toward involvement, which is enough to make the downstream implications worth taking seriously.
Implication one: the abuse-liability question comes back
Ketamine is a Schedule III controlled substance with well-documented recreational use and dependence potential. The field has generally treated its antidepressant mechanism as separate from that liability: the drug helps depression through glutamate, the argument goes, and the abuse risk is a separable property. If the antidepressant effect runs in part through the mu-opioid receptor, that separation weakens. The mechanism that produces the benefit may be entangled with the mechanism that produces the risk, which is precisely what Schatzberg warned in 2018. For a treatment increasingly delivered through lightly regulated infusion clinics on repeat schedules, an opioid-linked mechanism is not a footnote. It bears on tolerance, on dependence, and on how often the drug should be given.
Implication two: the esketamine franchise was positioned as not-an-opioid
The commercial stakes concentrate in esketamine, marketed by Johnson and Johnson as Spravato, the approved and widely used psychiatric product built on this chemistry. Esketamine is positioned and regulated as a novel glutamatergic agent, dispensed under a risk evaluation and mitigation strategy that addresses dissociation, sedation, and abuse potential, and its entire differentiation from the opioids it is meant to help patients avoid rests on being mechanistically distinct from them. Evidence that the underlying antidepressant effect involves the opioid system complicates that framing. It does not undo the approval or the efficacy data, and the read-across from racemic ketamine studies to esketamine is itself debated, which is the careful caveat. But it sharpens the long-term dependence question for a product taken chronically, and it does so against the backdrop of an opioid crisis that makes any opioid linkage commercially and politically sensitive.
Implication three: the next generation may not be able to remove the opioid part
A wave of companies is developing rapid-acting antidepressants pitched as delivering ketamine-like benefit without the dissociation and abuse risk, the clean version of the idea. The opioid evidence introduces a development risk for that strategy. If the antidepressant effect depends in part on mu-opioid engagement, then a molecule engineered to strip out the opioid activity may strip out the efficacy along with it. The buprenorphine study points the opposite direction, toward deliberately leaning into the opioid mechanism to extend the benefit, which is a coherent strategy precisely because it accepts what the clean-molecule programs are trying to design away. The two approaches cannot both be right about how much the opioid system matters, and the June papers tilt toward the side that says it matters.
The cautionary parallel
The clearest warning sits in the review’s own comparison. Tianeptine, an antidepressant available in Europe, was long thought to work through serotonin, and was later shown to be a mu-opioid receptor agonist that requires the receptor to produce its antidepressant effect. Tianeptine also carries documented misuse and abuse potential, to the point that it is sold illicitly in the United States. It is a worked example of an antidepressant whose efficacy and whose abuse liability turned out to share a mechanism. The concern with ketamine is that it belongs in the same category, and that the field has been slow to say so because the alternative story, a clean glutamatergic antidepressant, is more comfortable.
The takeaway
The mechanism debate is no longer academic. Whether ketamine’s antidepressant effect engages the opioid system bears directly on how it should be scheduled and monitored, on how esketamine is positioned and used over the long term, and on whether the next generation of rapid-acting antidepressants can deliver the benefit without the liability. The June issue does not settle the question. It moves the weight of evidence toward opioid involvement, and in doing so it makes the field’s preferred description, an antidepressant that has nothing to do with opioids, increasingly difficult to maintain.