Which Cannabis Compounds Have Been Tested for Gut Problems, and What the Trials Found About Symptoms Versus Inflammation

Across four randomized placebo-controlled trials in Crohn’s disease and ulcerative colitis, cannabis and cannabis oil improved symptom scores and quality of life while leaving endoscopic scores, C-reactive protein, and fecal calprotectin unchanged. In the 2021 Crohn’s trial published in the Journal of Crohn’s and Colitis, the activity index fell from 282 to 166 on cannabis oil versus 264 to 237 on placebo, while the endoscopic score moved from 10 to 7 on cannabis and 11 to 8 on placebo, a difference of nothing.

Symptom relief without measurable change in inflammation is a specific and clinically dangerous combination in inflammatory bowel disease, where ongoing inflammation causes structural damage whether or not the patient feels it.

A patient with Crohn’s disease who feels substantially better on cannabis may reasonably conclude that the disease is improving and reduce or stop other treatment. The trial data say that assumption is not supported. Symptoms and inflammation can move independently.

The second reason is that the gut is where cannabis produces one of its most distinctive harms. Cannabinoid hyperemesis syndrome presents as a gastrointestinal complaint, is frequently worked up as one, and is made worse by the treatment patients instinctively reach for.

Timna Naftali and colleagues in Israel have run most of the controlled work here, and the pattern across their trials does not vary.

The 2013 study in Clinical Gastroenterology and Hepatology randomized 21 patients with a Crohn’s Disease Activity Index above 200 who had failed steroids, immunomodulators, and anti-tumor necrosis factor agents. They smoked cannabis cigarettes containing 115 mg of THC twice daily, or placebo cigarettes made from flowers with the THC extracted, for eight weeks. The primary endpoint, complete remission defined as an index below 150, was reached by 5 of 11 on cannabis and 1 of 10 on placebo, which was not statistically significant at p = 0.43. Clinical response, defined as a fall of more than 100 points, occurred in 10 of 11 versus 4 of 10, p = 0.028. Three patients came off steroids. The trial is frequently cited as showing cannabis induces remission in Crohn’s disease. Its own authors state in the abstract that the primary endpoint was not achieved.

The 2021 trial in the Journal of Crohn’s and Colitis is the more informative one because it measured the gut directly. Fifty-six patients received oral cannabis oil containing 160 mg per millilitre cannabidiol and 40 mg per millilitre THC, or placebo, for eight weeks. The activity index fell from 282 to 166 on cannabis and from 264 to 237 on placebo. Quality of life rose from 74 to 91 versus 74 to 75. The Simple Endoscopic Score for Crohn’s Disease went from 10 to 7 on cannabis and 11 to 8 on placebo, p = 0.75. C-reactive protein and calprotectin were unchanged. The authors’ own conclusion is that until further studies are available, cannabis treatment in Crohn’s disease should be used only in the context of clinical trials.

In ulcerative colitis the picture repeated. A 2021 trial in PLOS ONE randomized 32 patients to cigarettes containing 0.5 g of dried cannabis flower with 80 mg THC or placebo for eight weeks. The Lichtiger disease activity index improved from 10.9 to 5 on cannabis versus 11 to 8 on placebo, significant between groups at p = 0.001, and quality of life improved substantially. The Mayo endoscopic score changed from 2.13 to 1.25 on cannabis and 2.15 to 1.69 on placebo, with a between-group p of 0.17. The authors wrote that the clinical benefits were not associated with significant anti-inflammatory improvement in the endoscopic score or laboratory markers.

The compound most heavily marketed for gut inflammation is the one with the clearest negative trials.

Naftali and colleagues randomized 20 patients with active Crohn’s disease to oral cannabidiol 10 mg twice daily or placebo for eight weeks, reported in Digestive Diseases and Sciences in 2017. The activity index fell from 337 to 220 on cannabidiol and from 308 to 216 on placebo. There was no difference. The paper’s title states the conclusion plainly: low-dose cannabidiol is safe but not effective. The authors note the dose may have been too low, which is a fair caveat and remains untested at higher doses in this population.

Peter Irving and colleagues ran a larger multicentre trial of a cannabidiol-rich botanical extract in ulcerative colitis over 10 weeks, published in Inflammatory Bowel Disease in 2018. The primary endpoint was negative: remission at end of treatment was 28 percent on extract and 26 percent on placebo. Per-protocol analyses of partial Mayo score and patient-reported outcomes favoured the extract, but a per-protocol analysis was needed because patients tolerated the extract poorly, taking on average one-third fewer capsules than the placebo group. Many of the adverse events were attributed to the THC content of the extract.

A trial where a third of the intended dose does not get taken is telling you something about tolerability, and a secondary analysis restricted to the people who managed the full dose is not the same evidence as a positive primary endpoint.

The strongest gastrointestinal evidence for cannabinoids is not about inflammation at all. It is about nausea.

The 2015 JAMA systematic review by Whiting and colleagues found that compared with placebo or active comparators, cannabinoids produced a greater average number of patients with a complete nausea and vomiting response, 47 percent versus 20 percent, with an odds ratio of 3.82 (95 percent CI 1.55 to 9.42) across three trials. The 2026 JAMA review reports a pooled standardized mean difference of -0.29 (-0.39 to -0.18) for prescribed cannabinoids such as dronabinol and nabilone against placebo or active comparators in nausea and vomiting from various causes. Small, real, and the basis for the existing regulatory approvals.

Gastroparesis produced a more surprising result. Ting Zheng and colleagues at the Mayo Clinic randomized 44 patients with idiopathic or diabetic gastroparesis to pharmaceutical cannabidiol escalated to 20 mg per kilogram per day, or placebo, for four weeks, reported in Clinical Gastroenterology and Hepatology in 2023. Cannabidiol reduced the total Gastroparesis Cardinal Symptom Index score (p = 0.008), vomiting episodes per 24 hours (p = 0.006), and overall symptom severity (p = 0.034), and improved tolerance of liquid nutrient intake. It also slowed gastric emptying further. Symptom relief occurred despite the physiology moving in the wrong direction, which is a useful reminder that gastroparesis symptoms and gastric emptying rates are loosely coupled.

THC does the same thing. A double-blind randomized study in Alimentary Pharmacology and Therapeutics gave 13 healthy volunteers THC at 10 mg per square metre and found that gastric emptying of solid food slowed substantially, with 73.9 percent of the isotope still in the stomach at 120 minutes versus 45.6 percent on placebo. Anyone with pre-existing delayed emptying should know that THC is likely to slow it further.

Irritable bowel syndrome is one of the most common reasons patients try cannabis for the gut and one of the least supported by controlled data.

The most rigorous work comes from the Mayo Clinic group. In a randomized trial in Neurogastroenterology and Motility, 36 volunteers with diarrhea-predominant irritable bowel syndrome received dronabinol at 2.5 mg or 5 mg twice daily or placebo for two days, with transit measured by validated scintigraphy. There was no overall treatment effect on gastric, small bowel, or colonic transit. A genotype-dependent signal appeared in carriers of a particular CB1 receptor variant, which is an interesting lead and not a treatment recommendation.

For cannabigerol, beta-caryophyllene, and the other compounds routinely described as gut anti-inflammatories in product marketing, the situation is simpler. Beta-caryophyllene is a genuine selective CB2 receptor agonist, demonstrated by Gertsch and colleagues in Proceedings of the National Academy of Sciences in 2008 through receptor binding, human monocyte assays, and a mouse inflammation model. That is real pharmacology. It is not a human gastrointestinal trial, and no such trial has been published.

Cannabigerol has recently acquired a small number of human studies in anxiety and tolerability, none of them in gastrointestinal disease. Describing what these compounds do to a human gut is describing something that has not been measured.

No page about cannabis and the gastrointestinal tract is complete without the syndrome cannabis produces there.

Cannabinoid hyperemesis syndrome is cyclic nausea and vomiting in regular cannabis users, characteristically relieved by hot showers or baths. Joseph Habboushe and colleagues surveyed emergency department patients at a New York public hospital who reported smoking cannabis at least 20 days per month. Among 155 such patients, 32.9 percent (95 percent CI 25.5 to 40.3) met the study’s symptom criteria for the syndrome. Extrapolated nationally, the authors estimated roughly 2.75 million Americans might experience something consistent with it annually.

The clinical trap is structural. The syndrome looks like a primary gastrointestinal disorder, so it gets an endoscopy, imaging, and a negative workup. Meanwhile the patient, reasonably, increases cannabis use to control nausea, because cannabis is the thing that has controlled nausea before. That escalation worsens the cycle.

Treatment evidence is thin. A 2024 systematic review in Academic Emergency Medicine examined seven studies and 492 patients and found mixed evidence for topical capsaicin and possible benefit from dopamine antagonists such as haloperidol and droperidol, while calling for methodologically rigorous trials. The only intervention with consistent effect is stopping cannabis, which is the part patients find hardest to accept and which resolves the syndrome when they do.

By the standards of cannabis research, the inflammatory bowel disease literature is unusually good. Four randomized placebo-controlled trials, prespecified primary endpoints, objective outcome measures including endoscopy and fecal calprotectin, and authors who reported negative primary endpoints plainly rather than burying them. The sample sizes are small, ranging from 20 to 56 patients, which limits precision but does not undermine the consistency of the pattern.

The nausea evidence is moderate quality and pooled across heterogeneous trials, and it underpins existing regulatory approvals. The irritable bowel evidence is essentially a null result from one well-designed transit study. The evidence for cannabigerol and beta-caryophyllene in human gastrointestinal disease is absent, not weak.

Every one of the inflammatory bowel trials ran for eight to ten weeks. That is long enough to see a symptom effect and arguably too short to see a mucosal healing effect, which in inflammatory bowel disease trials of conventional agents is usually assessed at least at week 12 and often later. The absence of endoscopic change may reflect duration as well as biology.

Blinding is again a problem. The two smoked-cannabis trials used THC-extracted placebo cigarettes, and participants receiving 80 to 115 mg of THC daily would have had little difficulty identifying their arm. Subjective activity indices such as the Crohn’s Disease Activity Index include patient-reported items and are vulnerable to that.

The gastroparesis trial ran for four weeks in 44 patients with a cannabidiol dose used in epilepsy, far above anything available over the counter. Its results should not be read as support for consumer cannabidiol products in gastrointestinal symptoms.

None of these trials show that cannabis or cannabidiol reduces intestinal inflammation in humans. Endoscopic scores, C-reactive protein, and fecal calprotectin were measured directly and did not change. That is a measured negative result, not an untested question.

No trial shows that cannabis prevents flares, reduces surgery or hospitalization, permits reduction of conventional therapy safely, or alters the long-term course of Crohn’s disease or ulcerative colitis. No published human trial tests cannabigerol or beta-caryophyllene for any gastrointestinal condition.

The gap between symptom scores and objective inflammation is one of the central problems in inflammatory bowel disease care generally, and it long predates cannabis. Patients with quiescent symptoms can have active inflammation, and patients with severe symptoms can have a healed mucosa. Modern treatment targets are defined by objective measures for exactly that reason.

Cannabis enters that context as a drug that reliably moves the subjective side and, on the evidence available, does not move the objective side. That makes it a plausible adjunct for symptom burden in a patient whose inflammation is being controlled by something else, and a poor candidate for anything resembling monotherapy.

It also raises a question the trials cannot answer. If a patient feels dramatically better and their calprotectin has not moved, has their quality of life improved or has their early warning system been muted? That is not a hypothetical concern in a disease where silent inflammation causes strictures and surgery.

This is the page where I most often disagree with how my own field talks. The 2013 Crohn’s trial is quoted everywhere as proof that cannabis induces remission, and the authors state in their own abstract that the primary endpoint was not achieved. That is not a subtle misreading. It is the opposite of what the paper says.

What I tell patients with inflammatory bowel disease is straightforward. Cannabis may genuinely make you feel better, and the trials support that. It has not been shown to reduce the inflammation that is doing the damage, and your gastroenterologist should keep measuring that independently of how you feel. If your symptom scores improve and your calprotectin does not, we have improved your day and not your disease.

And I ask about hot showers. Anyone with months of cyclic vomiting, a clean workup, and a daily cannabis habit gets that question, because the answer changes the entire diagnosis and nobody else in the chain has asked it.

For inflammatory bowel disease, cannabis improves symptoms and quality of life without measurable effect on endoscopic scores or inflammatory markers across four randomized trials, and low-dose cannabidiol alone has failed its primary endpoints twice. For nausea and vomiting, cannabinoids have modest pooled benefit and existing regulatory approval. For irritable bowel syndrome, a well-designed transit study found nothing. Anyone using cannabis for gut symptoms should keep objective inflammation monitored separately, and anyone with cyclic vomiting relieved by hot showers should be evaluated for cannabinoid hyperemesis syndrome.

The finding to carry forward is the dissociation between symptoms and inflammation, because it decides how cannabis should be positioned: possibly useful alongside disease-directed therapy, unsupported as a replacement for it. The finding not to carry forward is any claim that these trials demonstrated an anti-inflammatory effect in the human gut. They measured for one and did not find it.