A Phase II trial shows selonabant, a CB1 receptor antagonist, dramatically reduces the feeling of being high from oral THC when given at the same time as the cannabinoid. However, because the drug was administered alongside THC rather than after intoxication was established, the study does not answer whether selonabant can actually treat someone who is already experiencing a cannabis overdose in an emergency department.

Cannabis-related emergency department visits continue to climb as product potency rises across legal and illicit markets. Approximately 1.7 million annual ED visits in the United States are now associated with cannabis, and no approved pharmacological antidote exists. Current management is entirely supportive, relying on reassurance, benzodiazepines for severe agitation, and observation. A targeted CB1 receptor antagonist that could reverse or shorten acute cannabinoid intoxication would represent a paradigm shift in emergency management, analogous to what naloxone provides for opioid overdose. This trial is among the first prospective randomized evaluations of whether that goal is pharmacologically achievable.

Acute cannabinoid intoxication is an increasingly common emergency department presentation that currently lacks any targeted pharmacological intervention. Selonabant (ANEB-001) is a selective CB1 receptor antagonist being developed as a potential antidote for THC-mediated intoxication. The mechanistic rationale is straightforward: by competitively blocking CB1 receptors, selonabant should attenuate or reverse the downstream subjective and physiological effects of THC binding. This Phase II study enrolled 85 healthy, occasional cannabis users across two sequential parts, administering selonabant at doses ranging from 10 to 100 mg alongside standardized oral THC doses, then measuring pharmacodynamic responses over an eight-hour window using validated instruments for subjective intoxication, alertness, postural stability, and heart rate.

Results were striking for the primary subjective outcome: selonabant at 30 mg reduced “Feeling High” by 82.8% (P less than 0.0001) and at 10 mg by 80.4% (P less than 0.001) compared to placebo when coadministered with 21 mg oral THC. Alertness was preserved across all dose groups. Objective body sway improved significantly at the 10 and 30 mg doses. Notably, however, THC-induced tachycardia was not significantly attenuated at any dose level, and higher selonabant doses (50 and 100 mg) produced dose-dependent nausea and vomiting without additional efficacy. No clinically meaningful psychiatric adverse events were observed on single dosing. The authors acknowledge that a rescue-paradigm trial, testing the drug after intoxication is already established, is the essential next step before clinical translation can be pursued.

Selonabant Blocks the THC High. Now Comes the Hard Part.

Imagine a drug that could undo a cannabis overdose, one that erases the panic, the racing heart, the paralytic anxiety that sends tens of thousands of people to emergency rooms each year. That drug may be getting closer. But “getting closer” is doing a lot of work in that sentence. This Phase II trial demonstrates something genuinely valuable: selonabant, a CB1 receptor antagonist, can block approximately 80% of the subjective intoxication from oral THC when administered at the same time as the cannabinoid, at doses as low as 10 mg. The pharmacodynamic signal is convergent across multiple validated measures, including subjective “feeling high” scores, alertness preservation, and objective postural stability. The double-blind, pre-registered design is appropriate for this stage of development, and the finding that 10 mg outperforms higher doses on the tolerability-to-efficacy ratio is a practically important dose-selection result. This is real, encouraging pharmacology. Before any criticism, that contribution deserves acknowledgment.

The central problem, however, is not subtle. The study administered selonabant simultaneously with THC, meaning participants never became intoxicated before receiving the antagonist. This is like testing a fire extinguisher by spraying it on the wood before lighting a match. It tells you the chemical can prevent combustion, and that is useful information. But the patient who arrives at the emergency department did not take a preventive dose of anything. They are already burning. Whether selonabant can reverse established intoxication, shortening the duration or intensity of a high that is already peaking, is the entire clinical question, and this trial simply does not address it. The authors frame their results around “emergency treatment of acute cannabinoid intoxication” throughout the paper, and that framing stretches beyond what the coadministration design can support. There are additional gaps that matter in practice: THC-induced tachycardia, one of the most common reasons patients present to emergency departments, was not attenuated at any dose. The oral formulation would need to be swallowed by someone who may already be vomiting. And the healthy, screened, occasional-use volunteers in this trial bear little resemblance to the chronic users, adolescents, or synthetic cannabinoid-exposed patients who populate the acute intoxication literature.

What I would tell a patient is straightforward: this drug shows real promise in a laboratory setting, but it has not been tested in people who are already experiencing a bad cannabis reaction. What I would tell a colleague is that the 10 mg dose-finding result is the most practically useful takeaway, and that we need a rescue-paradigm trial before this changes anything we do in the emergency department. What I would tell a policymaker is that this is encouraging early-phase evidence that merits investment in a properly designed Phase III study, not a reason to fast-track approval. The broader lesson here is one that applies far beyond cannabinoid pharmacology: pharmacodynamic challenge studies that demonstrate blockade of an effect are necessary but not sufficient to establish clinical treatment efficacy. The timing, population, and administration context of real clinical use must be directly tested before translational claims are warranted.

This study sits at the early proof-of-concept stage of the drug development arc. It is not a pivotal efficacy trial and should not be treated as one. The pharmacodynamic signal is robust enough to justify advancement to Phase III, but the evidence base for clinical decision-making remains non-existent. No emergency physician should change current management practices based on these data. The study is most useful as a dose-finding exercise and a validation that the CB1 antagonist mechanism can produce large-magnitude attenuation of THC effects in humans.

From a pharmacological and safety standpoint, clinicians should be aware that the class history of CB1 antagonists includes serious psychiatric adverse events. Rimonabant, the first-generation CB1 inverse agonist developed for obesity, was withdrawn from the European market due to depression and suicidality in chronic use. Selonabant is being positioned for single-dose acute use, which carries a different risk profile, but single-dose psychiatric safety data from 85 healthy screened volunteers provide only a preliminary safety signal. The dose-dependent nausea and vomiting at 50 to 100 mg doses are relevant for emergency settings where patients may already be gastrointestinally distressed. The one concrete recommendation for clinicians tracking this area: watch for results from a rescue-paradigm trial in actual emergency department patients before drawing any conclusions about clinical utility.

This is a Phase II randomized, double-blind, placebo-controlled pharmacodynamic challenge study in healthy volunteers. In the evidence hierarchy, it occupies a position below Phase III efficacy trials and well below systematic reviews or meta-analyses of clinical outcomes. It is designed to establish whether a pharmacodynamic signal exists and to guide dose selection, not to demonstrate clinical efficacy. The single most important inference constraint is that the coadministration paradigm evaluates prevention of THC effects, not reversal of established intoxication.

Prior studies with first-generation CB1 antagonists or inverse agonists, including drinabant, surinabant, and rimonabant, demonstrated that prophylactic CB1 receptor blockade could attenuate the subjective effects of smoked cannabis in controlled settings. Selonabant’s results are consistent with and extend this earlier body of work by testing a newer compound specifically positioned for acute intoxication treatment rather than chronic appetite suppression. The magnitude of effect reported here, particularly the 80% reduction in “Feeling High” at 10 mg, appears larger than some prior challenge studies, though direct cross-study comparisons are complicated by differences in THC route, dose, and antagonist pharmacology. This trial confirms the mechanistic viability of CB1 antagonism for modulating acute THC effects but does not yet advance the evidence beyond the prevention paradigm that characterized all prior work in this space.

The most consequential analytical choice was the log-transformation of VAS “Feeling High” scores with an added 2 mm constant, which converts results into percentage reductions rather than absolute millimeter differences. This is a reasonable approach to handle the skewed distribution of VAS data, but it amplifies the apparent effect size when baseline values are moderate. Reporting absolute VAS differences alongside percentage reductions would provide a more balanced picture. Additionally, the absence of multiple comparison corrections across the many secondary endpoints means some statistically significant secondary findings may be false positives. A pre-specified hierarchical testing procedure or false discovery rate correction could have reduced the number of endpoints reaching nominal significance, though the primary outcome results are robust enough that corrections would be unlikely to overturn the central pharmacodynamic finding.

The most likely overinterpretation is concluding that selonabant is an effective treatment for acute cannabinoid intoxication. The trial tested whether the drug can prevent THC effects when given simultaneously, not whether it can reverse effects that are already established. These are pharmacologically and clinically distinct questions. A related misreading involves equating an 80% reduction in “Feeling High” with resolution of 80% of intoxication symptoms. “Feeling High” is one subjective metric on one visual analogue scale; THC-induced tachycardia was not attenuated, and the broader symptom burden of clinical intoxication, including anxiety, psychosis, and vomiting, was not assessed in this population. Readers should also avoid comparing dose-response results across Parts A and B, which used different THC doses and different sample sizes, making direct comparisons between dose arms across parts statistically invalid.

This Phase II trial establishes a robust pharmacodynamic signal that selonabant can block the subjective intoxicating effects of oral THC when given simultaneously, and identifies 10 mg as a promising dose that balances efficacy and tolerability. It does not establish clinical efficacy for treating patients who are already intoxicated, does not address THC-induced tachycardia, and does not provide evidence relevant to chronic users, pediatric populations, or synthetic cannabinoid exposures. For current clinical practice, management of acute cannabis intoxication remains supportive.

Is selonabant available as a treatment for cannabis overdose right now?

No. Selonabant is an investigational drug that has not been approved by any regulatory agency for any indication. It is currently in clinical development, and the results of this Phase II trial, while promising, represent an early stage of testing. Additional trials in actual emergency department patients are needed before it could be considered for clinical use.

Does this study prove the drug works as a cannabis antidote?

Not quite. The study shows that selonabant can prevent the feeling of being high when taken at the same time as THC. However, in real emergencies, people are already intoxicated when they seek help. Whether the drug can reverse an intoxication that is already underway has not yet been tested, and that distinction is clinically critical.

Were there safety concerns with this drug?

At lower doses (10 and 30 mg), selonabant was generally well tolerated. Higher doses (50 and 100 mg) caused nausea and vomiting in some participants without providing additional benefit. No psychiatric side effects were detected in this single-dose study, but the class of CB1 antagonist drugs has a history of psychiatric adverse effects with chronic use, so longer-term monitoring will be important in future trials.

Would this drug work for edible cannabis overdoses specifically?

The study used oral THC tablets, which somewhat mirrors edible cannabis exposure in terms of absorption timing. However, the drug was given at the same time as the THC, not hours later when edible effects typically peak and patients become distressed. Additionally, the drug did not reduce THC-induced rapid heart rate, which is a prominent concern in many edible overconsumption cases.

Have thoughts on this? Share it: