Clinical Insight | CED Clinic

A new randomized, controlled trial from the National Institute on Aging found that activating cannabinoid receptors with nabilone significantly improved insulin sensitivity in healthy humans, while blocking the CB1 receptor reduced it. Neither compound affected how much insulin the pancreas produced. For patients navigating cannabis alongside metabolic concerns, this adds controlled human evidence that the endocannabinoid system is not metabolically neutral.

Millions of Americans use cannabis while also managing blood sugar, prediabetes, or metabolic syndrome, yet clinical guidance on the intersection has been shaped more by assumption than controlled human data. This trial, funded by the National Institute on Aging and using the most rigorous metabolic methodology available, establishes that cannabinoid receptor activity has a measurable, directional effect on insulin sensitivity in the human body. That is not a small finding. It means every clinical conversation about cannabis and metabolic health now has a firmer mechanistic foundation to stand on.

The endocannabinoid system is known to influence energy balance, feeding behavior, and adipose tissue function, but its acute effects on insulin sensitivity in humans have been difficult to measure cleanly in real-world settings. Researchers at the NIA designed a four-way crossover trial in 21 healthy men to address this gap directly. Each participant underwent sequential hyperglycemic and euglycemic-hyperinsulinemic clamps after receiving nabilone (2 mg), a low-dose CB1 antagonist CP-945,598 (15 mg), a high-dose CB1 antagonist (45 mg), or placebo on separate visits at least six weeks apart. Circulating endocannabinoids and N-acylethanolamines were tracked throughout each session.

The primary finding was clear and directional: nabilone significantly improved insulin sensitivity during the euglycemic clamp, while high-dose CB1 blockade significantly reduced it. Neither compound affected pancreatic insulin secretion during the hyperglycemic phase. Elevated insulin levels suppressed circulating anandamide in a saturable, insulin-dependent pattern, while 2-arachidonoylglycerol (2-AG) was unaffected. The study population was limited to healthy, non-obese men, which constrains direct translation to patients with established metabolic disease, and the authors note that longer-term and population-specific trials are needed before clinical recommendations can be drawn.

The Body Already Knew: Why the Endocannabinoid System Was Never Just About Appetite

Every week in my practice I have a version of the same conversation. A patient managing prediabetes or metabolic syndrome asks whether their cannabis use is helping or hurting their blood sugar. Until recently, I could offer a reasonably honest answer drawn from mechanistic science and observational data, but I had to acknowledge that controlled human evidence was thin. This study changes that picture in a meaningful way. Not by giving us a prescription protocol. But by confirming what the biology was already suggesting: cannabinoid receptor activity is metabolically consequential in the human body, and the direction of the effect depends on whether you are activating or blocking the receptor.

The methodology here matters enormously and I want to spend a moment on it, because it is what makes this study worth reading carefully. The investigators used sequential hyperglycemic and euglycemic-hyperinsulinemic clamps. That is not a phrase to skim past. The clamp is the gold standard for isolating insulin secretion from insulin action. It lets you answer a question that you simply cannot answer from observational data or even from most standard clinical trials: does this compound change how much insulin the body makes, or does it change how well the body responds to insulin that is already there? This study answered that cleanly. Nabilone, acting as a cannabinoid receptor agonist, improved insulin sensitivity. It did not do so by changing how much insulin was secreted. The pancreas behaved the same way across conditions. What changed was the body’s response to insulin during the euglycemic phase. That is a specific and mechanistically meaningful finding.

The CB1 antagonist finding deserves equal attention, and perhaps more caution. At high doses, blocking CB1 reduced insulin sensitivity in healthy men. This reinforces what the clinical story of rimonabant already told us at a larger scale: the endocannabinoid system, and CB1 in particular, is not a straightforward target for metabolic intervention. Rimonabant was pulled from the European market in 2008 due to severe psychiatric side effects, and the CB1 antagonist class has largely stalled, though peripherally restricted versions are being revisited in current pharmaceutical research. The NIA data add a layer of concern to that revisitation: even in a healthy, non-psychiatric population, high-dose CB1 blockade measurably impairs the body’s ability to use insulin. That should inform how carefully any clinical application of this class is pursued.

The anandamide finding is one of the more underappreciated details in this paper. Elevated insulin during the clamps suppressed circulating anandamide in a saturable, dose-dependent way. The other primary endocannabinoid, 2-AG, was not affected. This tells us something real about the bidirectional relationship between the endocannabinoid system and insulin signaling: it is not symmetric. Anandamide and 2-AG, despite both being endocannabinoids and both acting on CB1 and CB2 receptors, do not respond to the same metabolic signals in the same way. That specificity is actually useful. It suggests that the ECS is not just a general metabolic regulator but has distinct physiological lanes, and that research treating all endocannabinoids as interchangeable is likely missing important signal.

What does this mean for the patient sitting across from me who uses cannabis and is managing blood sugar? I want to be precise here. This study was conducted in 21 healthy, non-obese men under tightly controlled experimental conditions. Nabilone is a synthetic, orally administered cannabinoid agonist dosed at 2 mg. Most of my patients are not using nabilone. They are using flower, or a vape, or a tincture, often at THC exposures that are hard to quantify precisely. The mechanistic principle, that CB1 agonism supports insulin sensitivity, may well apply more broadly, but we cannot assume it does. The dose, the delivery route, the patient’s metabolic baseline, their existing medications, all of these factors would modify the effect in ways this study was not designed to test.

What I can say with confidence is that this study advances the scientific foundation for a conversation that I think clinicians need to be having more deliberately. Cannabis is not just an appetite stimulant. The endocannabinoid system is woven into the regulatory biology of glucose and energy homeostasis in ways that are clinically relevant and increasingly measurable. Pretending that cannabis use is metabolically inert, in either direction, is no longer defensible from an evidence standpoint. The honest answer for my patients right now is that the physiology suggests cannabinoid activity affects insulin sensitivity, that the direction appears to favor agonism over antagonism, and that the rigorous, patient-population trials needed to translate this into clinical guidance do not yet exist. That is a more useful answer than either dismissal or false confidence. And it is the one this study earns.

This study sits at an important inflection point in endocannabinoid system research. Preclinical evidence linking CB1 receptor activity to insulin sensitivity has accumulated for more than two decades across adipose, hepatic, and skeletal muscle models. What has been missing is controlled human evidence using methodology rigorous enough to isolate the specific mechanism. Prior human data on cannabis and metabolic outcomes has largely come from epidemiological studies noting that cannabis users tend to have lower fasting insulin and BMI, a suggestive but confounded signal. The NIA clamp study is a meaningful step toward mechanism-level confirmation in humans, building on earlier work by Engeli et al. and by the rimonabant trials that implicated CB1 in adipogenesis and energy balance.

From a pharmacological standpoint, patients using cannabis alongside metformin, GLP-1 receptor agonists, or insulin should flag this to their prescribing clinician. While this study does not establish an interaction risk, the overlapping physiological terrain, specifically insulin signaling and glucose disposal, warrants awareness. Patients with insulin-dependent diabetes require particular attention given the insulin sensitivity effects demonstrated here. The concrete recommendation is practical: if you are using cannabis and managing blood sugar with medications, document your use and bring the conversation to your physician before assuming the combination is neutral in either direction.

This is a randomized, controlled, within-subject crossover physiological trial, a high-quality design for mechanistic questions in healthy volunteers. The clamp methodology is the gold standard for isolating insulin secretion from insulin sensitivity. The primary constraint on inference is the population: 21 healthy, non-obese men are not representative of the patients most likely to encounter cannabis-metabolic interactions in clinical practice, and no conclusions about long-term effects, therapeutic dosing, or patient-population outcomes can be drawn from this data alone.

Epidemiological studies have consistently noted that cannabis users have lower fasting insulin and reduced insulin resistance compared to non-users (Rajavashisth et al., 2012; Penner et al., 2013), a pattern observed even after controlling for BMI and dietary factors. The NIA trial adds the mechanistic layer those observational studies could not provide: a receptor-level, controlled demonstration that CB1 agonism improves insulin action in humans. This builds on the rimonabant trials, which showed that CB1 blockade produced weight loss and metabolic improvement in obese patients but caused serious psychiatric harm, confirming that the CB1 system is genuinely metabolically active while illustrating the risks of systemic blockade.

The finding that insulin suppresses anandamide but not 2-AG extends prior animal work on differential endocannabinoid tone regulation and suggests that translational research treating AEA and 2-AG as interchangeable metabolic markers may have been missing important physiological distinctions. This specificity is likely to become more important as the field moves toward targeted ECS pharmacology.

The crossover design with participants serving as their own controls is analytically strong for detecting within-subject effects, but the small N limits power for subgroup analyses and for detecting interactions between cannabinoid treatment and individual metabolic variability. A parallel-arm design with larger enrollment would have allowed for more robust covariate adjustment, particularly for baseline endocannabinoid tone, which likely varies meaningfully between individuals and could modify responsiveness to CB1 agonism or antagonism.

The study used nabilone as the agonist, not THC or plant-derived cannabinoids. Nabilone has a distinct pharmacokinetic profile and receptor binding pattern. An equivalent study using botanical THC preparations at comparable receptor exposure could produce meaningfully different results, and the gap between synthetic agonist data and real-world cannabis use remains an important limitation on direct clinical extrapolation.

The most foreseeable overreach from this study is the claim that cannabis improves blood sugar control and can serve as a treatment for insulin resistance or type 2 diabetes. This study establishes a receptor-level mechanistic effect in healthy volunteers under controlled conditions. It does not test any cannabis product. It does not include patients with metabolic disease. It does not measure long-term outcomes. Any headline framing this as evidence that cannabis treats diabetes exceeds what the data can support.

The inverse misreading, that the CB1 antagonist finding validates blocking the endocannabinoid system for metabolic benefit, is equally unwarranted. The reduced insulin sensitivity seen with CB1 blockade in this healthy population adds to the existing body of evidence that systemic CB1 antagonism carries real physiological costs, not just psychiatric ones.

This NIA trial establishes that cannabinoid receptor agonism improves insulin sensitivity and CB1 antagonism reduces it in healthy humans, using the most rigorous metabolic methodology available. It does not yet tell us how these effects translate to patients with metabolic disease, which cannabis products or doses would produce equivalent effects, or whether the benefits would persist over time. For clinical practice today, it means the endocannabinoid system belongs in the metabolic health conversation, but not yet as a prescriptive tool.

Does this mean cannabis can help control my blood sugar?

Not on its own, and not yet as a prescription strategy. This study shows that activating cannabinoid receptors improved insulin sensitivity in healthy volunteers under controlled lab conditions. It does not test cannabis products in patients with diabetes or prediabetes, and no one should adjust medications or dietary management based on this finding alone.

I use cannabis and I have prediabetes. Should I be concerned or encouraged?

Neither panic nor premature optimism is warranted. This study adds to a body of evidence suggesting the endocannabinoid system interacts meaningfully with insulin signaling, which is a reason to bring your cannabis use into the conversation with your physician rather than keeping the two tracks separate. It is not a reason to change anything unilaterally.

What is the endocannabinoid system and why does it affect metabolism?

The endocannabinoid system is a signaling network built into human physiology that regulates a wide range of functions including appetite, mood, inflammation, pain, and as this study confirms, insulin action. CB1 receptors, which respond to both the body’s own endocannabinoids and to THC from cannabis, are present in the liver, fat tissue, skeletal muscle, and pancreas, all tissues that play key roles in how the body manages blood sugar.

Is nabilone the same as cannabis?

No. Nabilone is a synthetic cannabinoid agonist that activates the same receptors as THC but has a different chemical structure, pharmacokinetic profile, and clinical effect pattern. It is FDA-approved for chemotherapy-induced nausea. The mechanistic principles from nabilone research may apply to botanical cannabis, but the two should not be treated as equivalent for purposes of dosing or clinical effect.

Should I tell my doctor I use cannabis if I am on diabetes medications?

Yes, absolutely. This study is one more reason why. If cannabis activity affects insulin sensitivity through cannabinoid receptor pathways, and you are also taking medications that work on blood sugar through other pathways, your physician needs the full picture to make informed decisions with you. Disclosing cannabis use is always in your clinical interest.

• Study population limited to 21 healthy, non-obese men; results cannot be directly extrapolated to women, older adults, or patients with existing metabolic disease
• Nabilone is a synthetic cannabinoid and is not equivalent to botanical cannabis products in pharmacokinetics, receptor binding, or clinical effect profile
• The crossover design controls well for individual variation but has insufficient power for subgroup analysis or for detecting modifier effects of baseline endocannabinoid tone
• Single-dose, acute protocol: no long-term metabolic outcomes measured; no data on tolerance, chronic effects, or dosing optimization

1. Chia CW, Tang EY, Cutler RG, et al. Acute metabolic effects of cannabinoid receptor modulators during sequential hyperglycemic, euglycemic-hyperinsulinemic clamps in healthy individuals. Am J Physiol Endocrinol Metab. 2026;330(5):E659-E674. doi:10.1152/ajpendo.00391.2025 | PMID: 41869774
2. Rajavashisth TB, Shaheen M, Norris KC, et al. Decreased prevalence of diabetes in marijuana users. PLoS One. 2012;7(1):e30198.
3. Penner EA, Buettner H, Mittleman MA. The impact of marijuana use on glucose, insulin, and insulin resistance among US adults. Am J Med. 2013;126(7):583-589.
4. Pi-Sunyer FX, Aronne LJ, Heshmati HM, et al. Effect of rimonabant, a cannabinoid-1 receptor blocker, on weight and cardiometabolic risk factors in overweight or obese patients. JAMA. 2006;295(7):761-775.

Have thoughts on this? Share it: