Review2008
The diverse CB1 and CB2 receptor pharmacology of three plant cannabinoids: delta9-tetrahydrocannabinol, cannabidiol and delta9-tetrahydrocannabivarin
Pertwee et al. · British Journal of Pharmacology
receptor-pharmacologycomparisonTHCCBD
Cannabinoid pillar · varin series
Tetrahydrocannabivarin (THCV).
The propyl homolog of THC — and not, as almost everyone assumes, shortened THC. It is sold as diet weed. No published study reports an appetite result in its favour; the trial in patients missed its own primary endpoint; and a 207-patient Phase 2 that would have settled the question completed in 2016 and was never published. Here is that record, in full.
Registry-verified·21 studies·updated 2026-08-09·What's this?
Origin & biosynthetic role
THCV is not shortened THC.
The commonest misconception about THCV is that the plant makes THC and then trims two carbons off it. It does not. THCV comes from its own acid, on a parallel line: the varin precursor CBGVA is converted by the same enzyme family that makes THC's acid, giving THCVA, which loses CO₂ with heat to give neutral THCV. Step for step it is the same shape as the pentyl route from CBGA — a different starting acid, not a modified product.
CBGVA
varin precursor (C₃)
THCA synthase
enzyme
THCVA
varin branch acid
Δ9-THCV
heat removes CO₂
THE VARIN LINE — this page's compound. The varin step is inferred from the characterised pentyl reaction below; the enzymology the registry holds is the pentyl one.
CBGA
pentyl precursor (C₅)
THCA synthase
enzyme
THCA
branch acid
Δ9-THC
heat removes CO₂
THE PENTYL LINE — the verified biosynthesis the varin line parallels. Note what is NOT here: no arrow runs from Δ9-THC to Δ9-THCV. The two are siblings, not parent and child.
Why the parallel matters more than it sounds
Because “THC with a shorter tail” invites the conclusion that THCV is a milder THC, and the pharmacology says the opposite can be true. At the CB1 receptor THCV blocks the cannabinoids that THC activates — until the dose rises, at which point the direction reverses. A compound that can be THC's opposite or THC's imitator depending on how much you take is not a weaker version of anything. The receptor detail is in the mechanism section, at Researcher depth.
The knowledge boundary
Exactly where THCV's certainty ends.
THCV is marketed on claims its evidence does not carry. Here is the honest line — including the claims we deliberately refuse. Read the right-hand column carefully: it holds two different kinds of gap, and the difference between them is the whole discipline of this page.
Established
- ✓Chemical identityC₁₉H₂₆O₂ (≈ 286.4 g/mol) — the propyl (C₃) homolog of THC's pentyl (C₅) side chain. Two carbons shorter, on the same tetrahydrocannabinol skeleton.
- ✓It is not shortened THC — it has its own acidTHCV is not made by trimming THC. The plant runs a parallel line: the varin precursor CBGVA is converted by THCA synthase to THCVA, which loses CO₂ with heat to give neutral Δ9-THCV — step for step the same shape as CBGA → THCA → Δ9-THC. The varin branch is inferred from the characterised pentyl reaction; the enzymology of the pentyl step is what the registry holds.
- ✓Its cannabinoid-receptor directions are published — with a dose conditionThis is the correction: the directions exist and are quoted verbatim on this page. THCV antagonises cannabinoid agonists in CB1-expressing tissues, and in a live animal behaves as a CB1 antagonist or, at higher doses, as a CB1 agonist. The dose qualification is part of the finding, not a footnote to it.
- ✓It has been given to people, in registered controlled trialsNot many, and none large — but the record is real and it is published in full on this page: a 62-participant metabolic pilot, an interaction study in ten men, and a dose-ranging safety and pharmacokinetics programme running to 200 mg. What those trials found is the next column.
- ✓At higher doses it produces effects people can feelIn the dose-ranging safety study, 100 mg and 200 mg produced elevated ratings of “feel a drug effect” and “like the drug effect”, and the authors state that THC-like effects were observed at higher THCV doses. THCV is routinely described online as non-psychoactive; at the top of the tested range, in people, that description does not hold.
- ✓It can make a urine drug screen read positive for THCIn the same programme, almost every post-dose urine screen came back positive for THC — and a follow-up analysis found the majority of pooled samples flagged positive by a routine immunoassay while the reference method found no Δ9-THC-COOH at all. Documented false positives, in people, from a product that is not THC.
Not yet established
- —Any approved use for diabetes, weight loss or appetiteNot established. There is no approved THCV medicine in any jurisdiction, for any indication. Nothing on this page should be read as suggesting that THCV treats, prevents or manages type 2 diabetes, obesity or anything else.
- —The metabolic trial's own primary endpointNot established, and not for want of testing. The registered primary endpoint of the 62-participant trial was HDL cholesterol, and the paper reports “plasma HDL was unaffected”. That is a result against the endpoint the trial itself chose in advance, not a gap in the data.
- —Appetite suppression, in a personNot established, and this is the claim THCV is sold on. Appetite was a REGISTERED SECONDARY ENDPOINT of the metabolic trial, scored daily for thirteen weeks, and no appetite result appears in the published paper. Two studies that measured subjective ratings directly found no significant difference from placebo. There is no positive human appetite result to cite.
- —Which direction THCV takes at CB2Genuinely unresolved. One paper reports a CB2 partial agonist; another, from the same laboratory lineage, reports a CB2 antagonist. Both are published and neither has been retracted. We report the conflict and assert neither direction.
- —Where the CB1 dose flip sits in a personNot established. The reversal from antagonist to agonist is documented, and the doses at which it was measured are from mice given THCV intravenously. No equivalent threshold has been established for a person taking THCV by mouth, so no dose on this page is described as “low” or “high” for a human.
- —Long-term safety, and drug interactionsNot established. The longest published exposure is thirteen weeks in 62 people; the dose-ranging safety study gave single acute doses. No interaction study in people has been published, and nobody has followed anyone taking THCV for a year.
Three different states live in that right-hand column — and they are not interchangeable
Tested, and the endpoint came back null. The metabolic trial's HDL primary; the subjective ratings in both imaging studies. Those are results, not gaps. Never tested. Long-term safety; where the CB1 dose flip sits in a person. Nobody has run those studies at all. Tested, finished, and never published. A 207-patient Phase 2 completed a decade ago whose result no one outside its sponsor can read. That third state is the rarest and the most consequential, and this page gives it its own section rather than letting it hide inside “more research is needed”. An absence of evidence, a negative result and a withheld result are three different things, and our evidence ladder reserves Not-supported for evidence weighing against a claim — a judgement for the gated pipeline rather than for this page.
What has actually been tested in people
The human record, in full.
Two published trials have given THCV to people to test what it does — one metabolic, one an interaction study against THC. Each entry leads with what was actually tested and carries a paired shows / does not show that cannot be separated from it. The safety and pharmacokinetic programme is published further down, in its own section, because it tested tolerability rather than any therapeutic claim.
Off-registry — cited in prose, no evidence grade
These papers are real, checkable, and not graded entries in the CBG Atlas registry — the gated pipeline has not ingested them, so none carries a rung on our evidence ladder and nothing here is asserted on their authority. The badges name each study's design, which is not a grade and is deliberately rendered in the neutral chip. If and when these enter the registry through the gated pipeline, they will appear graded and cited like every other record.
2016
Efficacy and Safety of Cannabidiol and Tetrahydrocannabivarin on Glycemic and Lipid Parameters in Patients With Type 2 Diabetes: A Randomized, Double-Blind, Placebo-Controlled, Parallel Group Pilot Study
What was actually tested: A randomised, double-blind, placebo-controlled, parallel-group study in 62 people with non-insulin-treated type 2 diabetes, across FIVE arms for 13 weeks: CBD 100 mg twice daily; THCV 5 mg twice daily; a 1:1 CBD:THCV combination (5 mg/5 mg twice daily); a 20:1 combination (100 mg/5 mg twice daily); and matched placebo. ⚠ IT IS INDEXED IN PUBMED WITH THE MeSH TERM “PILOT PROJECTS”, AND THE WORD “PILOT” IS IN ITS OWN TITLE — twelve or thirteen people per arm. ⚠ ITS REGISTRATION (NCT01217112) LISTS THE CONDITION AS DYSLIPIDAEMIA IN TYPE 2 DIABETES, and enrolment required LOW HDL at screening: the trial was built around a lipid question, and the registration's brief summary states it was conducted by GW Pharma Ltd. ⚠ THE REGISTRATION ALSO DATES IT: the study ran from October 2010 to September 2012 and this paper appeared in 2016 — a four-year gap between the last participant and the publication.
⚠ THE PRIMARY ENDPOINT WAS NOT MET. The registered primary was change in HDL cholesterol, and the paper reports, verbatim, that “plasma HDL was unaffected”. THCV alone did significantly reduce fasting plasma glucose versus placebo (estimated treatment difference −1.2 mmol/L, P < 0.05) and produced statistically significant differences on pancreatic β-cell function (HOMA2), adiponectin and apolipoprotein A. ⚠ AND THE ABSTRACT CONTRADICTS ITSELF ON THE DIRECTION OF THOSE THREE: it describes them as “improved”, yet every one of the three estimated treatment differences it prints is NEGATIVE — while the trial's own registration states that for β-cell function and for apolipoprotein A an INCREASE is what indicates improvement. This page therefore states that those three moved significantly and does NOT state which way. ⚠ Verbatim, and routinely dropped: “None of the combination treatments had a significant impact on end points.” ⚠ And the CBD results are versus BASELINE, not versus placebo — the paper says so.
What it shows: That in a 62-person pilot, 5 mg of THCV twice daily for thirteen weeks lowered fasting plasma glucose relative to placebo in people with type 2 diabetes, and moved three other blood measures by a statistically significant amount.
What it does NOT show: That THCV treats diabetes, or affects HDL — the endpoint this trial pre-specified to answer its own question was HDL, and it did not move. It shows nothing about appetite: appetite was a registered secondary endpoint, scored daily for the whole thirteen weeks, and NO APPETITE RESULT IS REPORTED. It shows nothing about the CBD-plus-THCV products sold on the strength of it — every combination arm was null. And a change in a blood marker over thirteen weeks in twelve-odd people per arm is not a clinical outcome; nobody's diabetes was shown to be better managed.
Read it with this: ⚠ INDUSTRY AUTHORSHIP, AND A CONCLUSION WIDER THAN THE RESULT. Colin Stott is affiliated to GW Pharmaceuticals on the paper, and the registration records the study as conducted by GW Pharma Ltd. The abstract's closing sentence — that THCV “could represent a new therapeutic agent in glycemic control” — is a conclusion about a SECONDARY domain, drawn from a pilot whose PRIMARY endpoint was null, and CBG Atlas does not echo it. ⚠ Read this trial next to the section below it: the same sponsor then ran a 207-patient Phase 2 with HbA1c as the primary endpoint, and never published it.
Randomised, placebo-controlled · in people · PILOT · primary endpoint missedOff-registry · carries no evidence grade
2016
The effect of five day dosing with THCV on THC-induced cognitive, psychological and physiological effects in healthy male human volunteers: A placebo-controlled, double-blind, crossover pilot trial
What was actually tested: A within-subject, placebo-controlled, double-blind crossover PILOT in TEN MALE cannabis users with fewer than 25 lifetime use occasions. Each received 10 mg of oral pure THCV or placebo daily for five days, followed on the fifth day by 1 mg of intravenous Δ9-THC. ⚠ THE QUESTION THIS STUDY ASKED WAS WHETHER THCV BLUNTS THC — not whether THCV does anything on its own, and not anything metabolic.
THCV was well tolerated and subjectively indistinguishable from placebo on its own. It inhibited the rise in heart rate that THC caused (p = 0.028), and nine of the ten participants rated THC as subjectively weaker or less intense under THCV (p = 0.011). ⚠ AND IT MADE ONE OUTCOME WORSE: “THCV in combination with THC significantly increased memory intrusions” (p = 0.031). ⚠ THE LIMIT IS IN THE ABSTRACT ITSELF: “THC did not significantly increase psychotic symptoms, paranoia or impair short-term memory, while still producing significant intoxicating effects.” The authors close by saying the findings need cautious interpretation because of the small sample and the absence of THC-induced psychotomimetic and memory-impairing effects, “probably owing to the choice of dose”.
What it shows: That five days of 10 mg THCV blunted the heart-rate rise from a small intravenous dose of THC in ten men, and that most of them found the THC weaker.
What it does NOT show: That THCV protects against THC's psychological effects. ⚠ THOSE EFFECTS LARGELY DID NOT OCCUR IN THIS STUDY — the THC dose did not significantly raise psychotic symptoms, paranoia or short-term memory impairment, so there was little there to protect against, and the authors say so. It also does not show THCV is benign in combination: memory intrusions went UP. And it says nothing whatever about metabolism, appetite or weight — this is an interaction pilot in ten men.
Read it with this: ⚠ THE STUDY MOST OFTEN CITED AS “THCV COUNTERACTS THC” CONTAINS A WORSENED OUTCOME IN THE SAME ABSTRACT. Reported here in full because a reference that publishes the heart-rate result and omits the memory-intrusion result is running a selection filter. Ten participants, all male, one dose level, one THC challenge.
Crossover pilot · in people · n = 10 · an INTERACTION study, not an efficacy trialOff-registry · carries no evidence grade
Primary · secondary · tertiary
Which THCV endpoint moved — and which one the trial existed to test?
A trial names one outcome, in advance, as the one that will answer its question. Everything else is secondary, and a positive further down the list does not fill the hole a missed primary leaves. Below is every outcome each trial registered, in the rank the registration gave it. The ranks are transcribed from ClinicalTrials.gov, not from the papers' own summaries — which is how the appetite endpoint below became visible at all.
Jadoon 2016 — THCV 5 mg twice daily, 13 weeks, n = 62 across five arms
Registration: NCT01217112 (study code GWMD1092). Registered conditions: DYSLIPIDAEMIA and type 2 diabetes. Enrolment required low HDL at screening (≤ 1.3 mmol/L in women, ≤ 1.2 mmol/L in men), so the population was selected for the lipid problem the primary endpoint measured. The registration's own summary states the study was conducted by GW Pharma Ltd; GWP42004 is THCV and GWP42003 is CBD. Results are posted to the registration. Funding: Industry. Colin Stott is affiliated to GW Pharmaceuticals on the paper; the registration names GW Pharma Ltd as the conducting company and Jazz Pharmaceuticals as the sponsor of record.
| Registered rank | Pre-specified outcome | What happened |
|---|---|---|
| PrimaryNot met | Change from baseline in mean serum HDL cholesterol concentration after 91 days (13 weeks) of treatment — the registration adds that an INCREASE indicates improvement | NOT MET. The paper states, verbatim, “although plasma HDL was unaffected”. The endpoint this trial was built around, in a population enrolled for having low HDL, did not move. |
| SecondarySignificant | Change from baseline in mean fasting glucose concentration | THCV alone reduced it versus placebo — estimated treatment difference −1.2 mmol/L, P < 0.05. Verb and sign agree here, so the direction is stated. |
| SecondarySignificant · direction unresolved | Change from baseline in mean insulin β-cell function measured by HOMA2 — the registration adds that an INCREASE indicates improvement | ⚠ SIGNIFICANT, DIRECTION UNRESOLVED. The paper says β-cell function was “improved” and prints ETD = −44.51 points, P < 0.01. A negative value is a DECREASE, and the registration says an increase is the improvement. The verb and the sign disagree; this page reports the significance and refuses to state the direction until the full text settles it. |
| SecondarySignificant · direction unresolved | Change from baseline in mean serum apolipoprotein A concentration — the registration adds that an INCREASE indicates improvement | ⚠ SIGNIFICANT, DIRECTION UNRESOLVED. Described as “improved”, reported as ETD = −6.02 μmol/L, P < 0.05. Same contradiction as the row above. |
| SecondarySignificant · direction unresolved | Adiponectin (reported in the paper among the endpoints THCV moved) | ⚠ SIGNIFICANT, DIRECTION UNRESOLVED. Described as “improved”, reported as ETD = −5.9 × 10⁶ pg/mL, P < 0.01. |
| SecondaryNot reported | Change from baseline in mean appetite, scored DAILY by participants on a 0–10 numerical rating scale for the whole 13 weeks | ⚠ REGISTERED, MEASURED EVERY DAY FOR THIRTEEN WEEKS — AND NO RESULT APPEARS IN THE PUBLISHED PAPER. This is the endpoint THCV's entire consumer reputation rests on, and it is the one the publication is silent about. |
| SecondaryNot reported | Change from baseline in mean glycated haemoglobin (HbA1c) concentration | Registered as a secondary. Not reported in the published abstract. ⚠ HbA1c became the PRIMARY endpoint of the 207-patient trial below — the one that was never published. |
| SecondaryNot reported | Change from baseline in mean body weight, BMI, waist measurement, waist-to-hip ratio, and total, visceral, subcutaneous and liver fat by magnetic resonance imaging | All registered as secondary outcomes and measured. None is reported in the published abstract, so this page states nothing about THCV and body weight or body fat in people. |
| SecondaryNull | All endpoints, in the two CBD + THCV COMBINATION arms | NULL, in the paper's own words: “None of the combination treatments had a significant impact on end points.” Every product that combines the two is sold against this sentence. |
| Tertiary / otherNot reported | Total and LDL cholesterol, HDL:LDL ratio, VLDL, triglycerides, apolipoprotein B, non-esterified fatty acids, fructosamine, OGTT glucose and insulin, fasting insulin, C-peptide, HOMA2-IR, insulin sensitivity, adverse events, and depression score (BDI-II) | Registered as secondary outcomes and not individually reported in the published abstract. Results are posted to the registration; nothing is stated here that this page has not read in a source it cites. |
The sentence that gets lifted out — and the rank it came from
The line that travels is “a clinical trial showed THCV lowers blood sugar in type 2 diabetes”. Fasting glucose is a SECONDARY endpoint, measured in roughly a dozen people on THCV, in a study whose own title calls it a pilot. The endpoint the trial pre-specified as the one that would answer its question — HDL, in a population enrolled for low HDL — was unaffected. And the appetite endpoint that would speak to why anyone buys THCV was registered, scored daily for thirteen weeks, and never reported.
GWP42004 Phase 2 — 2, 5 and 15 mg twice daily added to metformin, 12 weeks, n = 207
Registration: NCT02053272. A randomised, double-blind, placebo-controlled, parallel-group dose-ranging Phase 2 study of GWP42004 — THCV — as add-on to metformin in type 2 diabetes, across 25 sites in Romania and the United Kingdom. Started February 2014. STATUS: COMPLETED, with a primary completion date of FEBRUARY 2016. ⚠ NO RESULTS ARE POSTED TO THE REGISTRATION, AND NO PUBLICATION EXISTS. Searching PubMed for the compound code “GWP42004” returns ZERO records — checked 2026-08-23. Funding: Industry. Sponsor of record: Jazz Pharmaceuticals. GWP42004 is a GW Pharmaceuticals compound code, and the same code appears in the registration of the 62-patient pilot above.
| Registered rank | Pre-specified outcome | What happened |
|---|---|---|
| PrimaryNever published | Change from baseline to end of treatment in mean glycosylated haemoglobin (HbA1c) — the single measure that would establish whether THCV improves glycaemic control | ⚠ UNKNOWN, AND UNKNOWABLE FROM THE PUBLIC RECORD. The trial completed in February 2016. A decade later there are no posted results and no paper. Whatever this endpoint did, nobody outside the sponsor can say. This row is not an absence of research — it is an absence of PUBLICATION, and the two are entirely different things. |
| SecondaryNever published | Fasting plasma glucose; two-hour glucose and insulin on oral glucose tolerance testing; serum fructosamine; the number of participants reaching HbA1c below 7%; fasting insulin; HOMA2-IR; HOMA2-%B; pro-insulin; C-peptide | All registered. None published. ⚠ NOTE THE FIRST ONE: fasting plasma glucose is the endpoint the 62-person pilot moved, and the properly-powered test of it sits inside this unpublished trial. |
| SecondaryNever published | BMI; total cholesterol; HDL cholesterol; triglycerides; C-reactive protein; blood pressure | All registered. None published. HDL appears here as a SECONDARY, having been the pilot's failed PRIMARY. |
| SecondaryNever published | Incidence of adverse events; incidence of hypoglycaemic episodes; Beck Depression Inventory-II; Columbia-Suicide Severity Rating Scale; Cannabis Withdrawal Scale | ⚠ SAFETY OUTCOMES, REGISTERED AND UNPUBLISHED. 207 people took a cannabinoid twice daily for twelve weeks and the recorded safety data has never been made public. The depression and suicidality instruments are there because the drug class this compound sits beside — CB1 blockers — was withdrawn for psychiatric harm. |
| Tertiary / otherNever published | Diabetes Treatment Satisfaction Questionnaire; overall health visual analogue scale | Registered. Not published. |
The sentence that gets lifted out — and the rank it came from
There is no line to lift, and that is the point. Every confident THCV metabolic claim in circulation traces to the 62-patient PILOT above. The trial designed to test that pilot's hypothesis properly — four times the size, HbA1c as its primary endpoint, twelve weeks on top of metformin — finished ten years ago and was never published. A reference that reports the pilot and stays quiet about the unpublished Phase 2 gives readers the flattering half of a two-part story.
Why this is a table and not a paragraph
Because in a paragraph it does not survive. “THCV lowered blood sugar” and “the endpoint the trial pre-specified was HDL and it was unaffected” cannot sit in the same sentence without one of them winning, and the one that wins is always the shorter, happier one. Here the rank is a column, so a result cannot be lifted away from the rank it was registered under — and an endpoint that was registered but never reported gets a row of its own instead of vanishing. That is how the appetite endpoint, and the entire second trial, become visible.
The trial that was never published
207 patients. Twelve weeks. Finished in 2016. Never published.
This is the honest headline of THCV's metabolic story, and it is not a claim about what THCV does — it is a fact about what the public record contains. Every confident statement about THCV and blood sugar traces back to a 62-patient pilot whose own primary endpoint was null. The trial built to test that pilot properly exists, completed, and has never been reported.
- Registration
- NCT02053272
- Registered title
- A Randomised, Double Blind, Placebo Controlled, Parallel Group, Dose Ranging Study of GWP42004 as Add on to Metformin in the Treatment of Participants With Type 2 Diabetes
- Sponsor
- Jazz Pharmaceuticals (sponsor of record). GWP42004 is a GW Pharmaceuticals compound code for THCV, and the same code appears in the registration of the 62-patient pilot.
- Design
- Phase 2 — randomised, double-blind, placebo-controlled, parallel-group, dose-ranging
- Participants
- 207 participants with type 2 diabetes on stable metformin, at 25 sites in Romania and the United Kingdom. Doses of 2, 5 and 15 mg twice daily against placebo, for 12 weeks.
- Status
- COMPLETED. Started February 2014; primary completion February 2016; overall completion February 2016.
- Primary endpoint
- Change from baseline to end of treatment in mean glycosylated haemoglobin (HbA1c) — the standard measure of glycaemic control over months, and the endpoint a diabetes medicine has to move.
- Results posted
- None. The registration carries no results, a decade after completion.
- Publication
- ⚠ NONE. No paper reporting this trial has been published. A PubMed search for the compound code “GWP42004” returns ZERO records — checked 2026-08-23.
Why this matters more than any positive result
Because every confident metabolic claim made for THCV traces back to a 62-patient PILOT whose own primary endpoint was null, and this is the trial that was supposed to settle it: more than three times the size, with HbA1c — the endpoint that actually matters in diabetes — as its primary, run on top of standard first-line therapy. It finished. Its result is not public. That is the honest headline of THCV's metabolic story, and it is a stronger position than any claim: the platform that tells you what is missing is the one you can trust about what is there.
And what it does NOT mean
⚠ IT DOES NOT MEAN THE TRIAL FAILED. We do not know what it found, and this page does not guess. Trials go unpublished for many reasons — commercial decisions, programme discontinuation, restructuring — and inferring a negative result from silence would be exactly the kind of unfounded claim this platform refuses. What it means is narrower and firmer: THE LARGEST TEST OF THCV'S BEST-KNOWN CLAIM IS NOT IN THE EVIDENCE BASE, so any summary of THCV's metabolic evidence that does not say so is describing a subset and calling it the record.
The same programme published its small trial four years late, and its large one not at all
The 62-patient pilot ran from October 2010 to September 2012 and appeared in print in 2016 — a four-year gap between the last participant and the paper. The 207-patient Phase 2 started in February 2014 and completed in February 2016; a decade on, its registration still carries no results. Neither fact tells you what THCV does. Both tell you that the published THCV literature is a selected subset of the THCV evidence that exists, and that any summary which does not say so — including every consumer article written about this compound — is describing the selection and calling it the science.
The claim THCV is sold on
“Diet weed” has no human evidence behind it.
Not thin evidence. Not early evidence. No positive human result at all, from any published study, on appetite, food intake or body weight. That is an unusual thing to be able to say, so it is said with the studies attached rather than as an assertion — every claim below can be checked in the sources beneath it.
The four things the record actually contains
One. Appetite was a registered secondary endpoint of the trial in patients — participants scored it daily on a 0–10 scale for the whole thirteen weeks (NCT01217112) — and the published paper reports no appetite result. Two. A brain-imaging study measured how much twenty volunteers wanted and liked food after a THCV dose and found no significant differences in subjective ratings — while brain responses to both reward and aversion went up (Tudge 2015). Three. A second study in twenty volunteers reported no significant differences in the subjective experience (Rzepa 2016). Four. The mouse study the whole claim is traced to says, verbatim, that THCV “did not significantly affect food intake or body weight gain in any of the studies” (Wargent 2013). ⚠ Four sources, four directions of approach, and not one positive appetite result among them.
2015
Neural effects of cannabinoid CB1 neutral antagonist tetrahydrocannabivarin on food reward and aversion in healthy volunteers
What was actually tested: A within-subject, double-blind study in 20 healthy volunteers, each given a single 10 mg dose of THCV and placebo on two separate occasions in randomised order. Brain responses to rewarding stimuli (the sight and flavour of chocolate) and aversive ones (a picture of mouldy strawberries, a less pleasant strawberry taste) were measured by functional MRI. Volunteers also rated pleasantness, intensity and WANTING for each stimulus. ⚠ NOBODY'S FOOD INTAKE WAS MEASURED — this is a brain-imaging study, not an eating study.
⚠ THE SUBJECTIVE MEASURE WAS NULL, VERBATIM: “There were no significant differences between groups in subjective ratings.” The imaging did move, in BOTH directions: THCV increased brain responses to the chocolate stimuli in the midbrain, anterior cingulate cortex, caudate and putamen — AND increased responses to the aversive stimuli in the amygdala, insula, mid-orbitofrontal cortex, caudate and putamen. The authors' own framing is that THCV increased neural responding to rewarding AND aversive stimuli.
What it shows: That a single 10 mg dose of THCV changed how twenty people's brains responded to food images and tastes, turning up the response to pleasant and unpleasant stimuli alike.
What it does NOT show: That THCV suppresses appetite. The measures closest to appetite — how much people wanted the food, and how pleasant they found it — did not differ from placebo. ⚠ AND AN INCREASED RESPONSE TO FOOD REWARD IS NOT AN APPETITE-SUPPRESSING SIGNAL; if anything it points the other way, which is why this study is quoted for its conclusion rather than its result. Nothing was eaten and no weight was measured.
Read it with this: ⚠ THE PAPER'S OWN CONCLUSION IS THE PART THAT TRAVELS: it suggests “therapeutic activity in obesity, perhaps with a lowered risk of depressive side effects”. That is a hypothesis about a drug class, offered on the strength of a brain-imaging pattern in twenty people, against a null on everything they reported feeling. It is not an appetite finding and this page does not repeat it as one.
Randomised crossover · in people · n = 20 · brain imaging, not eatingOff-registry · carries no evidence grade
2016
The CB1 Neutral Antagonist Tetrahydrocannabivarin Reduces Default Mode Network and Increases Executive Control Network Resting State Functional Connectivity in Healthy Volunteers
What was actually tested: A randomised, within-subject, double-blind study in 20 healthy volunteers given a single 10 mg oral dose of THCV and placebo. Resting-state functional MRI with seed-based connectivity analysis in the amygdala, insula, orbitofrontal cortex and dorsomedial prefrontal cortex. Mood and subjective experience were measured before and after dosing with self-report scales.
⚠ THE SUBJECTIVE RESULT WAS NULL, AND THE AUTHORS EXPECTED IT TO BE: “Our results revealed, as expected, no significant differences in the subjective experience with a single dose of THCv.” Connectivity did change — reduced between the amygdala and the default mode network, increased between the amygdala and the dorsal anterior cingulate cortex, and increased between the dorsomedial prefrontal seed and the inferior frontal and medial frontal gyri.
What it shows: That a single 10 mg dose of THCV measurably altered resting brain connectivity in twenty healthy volunteers.
What it does NOT show: Anything about appetite, eating or weight — none was measured. And it confirms, in a second independent sample, that a 10 mg dose of THCV produces no detectable change in how a person feels. A connectivity pattern is a mechanistic observation, not a clinical outcome, and the paper's suggestion of “possible therapeutic activity for obesity” is a hypothesis, not a result.
Randomised crossover · in people · n = 20 · resting-state imagingOff-registry · carries no evidence grade
2013
The cannabinoid Δ9-tetrahydrocannabivarin (THCV) ameliorates insulin sensitivity in two mouse models of obesity
What was actually tested: Four oral dose-ranging studies in two mouse models of obesity — diet-induced and genetically obese (ob/ob) — running 30 to 45 days, plus cell work in insulin-resistant hepatocytes and myotubes. A CB1 inverse agonist was the positive control. Food and water intake, body weight gain, energy expenditure, glucose and insulin, lipids and liver triglycerides were all measured. MICE THROUGHOUT.
⚠ THIS IS THE SENTENCE THE “DIET WEED” CLAIM IS USUALLY TRACED TO, AND IT SAYS THE OPPOSITE, VERBATIM: “THCV did not significantly affect food intake or body weight gain in any of the studies, but produced an early and transient increase in energy expenditure.” What THCV did do was metabolic rather than appetitive: it dose-dependently reduced glucose intolerance in ob/ob mice, improved glucose tolerance and increased insulin sensitivity in diet-induced obese mice, and restored insulin signalling in the insulin-resistant cells — without consistently affecting plasma lipids.
What it shows: That in two mouse models of obesity, oral THCV improved measures of glucose handling and insulin sensitivity over 30 to 45 days.
What it does NOT show: ⚠ THAT THCV REDUCES FOOD INTAKE OR BODY WEIGHT — IN ANY OF THE FOUR STUDIES, IN MICE. The authors state that plainly, and the rise in energy expenditure they did see was early and transient. It also shows nothing about people: this is the preclinical work that preceded the human pilot, and the human pilot's primary endpoint was null.
Read it with this: ⚠ THE SAME PROGRAMME. C. G. Stott, a co-author here, is the author affiliated to GW Pharmaceuticals on the 2016 human pilot. This is the animal work behind that trial, and it is the animal work most often cited — usually for a weight-loss finding it explicitly does not contain.
Preclinical · MICE · four dose-ranging studies · not peopleOff-registry · carries no evidence grade
2017
Cannabis Edibles: Blood and Oral Fluid Cannabinoid Pharmacokinetics and Evaluation of Oral Fluid Screening Devices for Predicting Δ9-Tetrahydrocannabinol in Blood and Oral Fluid following Cannabis Brownie Administration
What was actually tested: A controlled oral cannabis administration study — participants ate cannabis brownies — with blood and oral fluid sampled to characterise cannabinoid pharmacokinetics and to evaluate roadside oral-fluid screening devices. Registered as NCT02177513. This is a forensic-toxicology study, and THCV appears in it as one of the minor cannabinoids measured.
⚠ THE MINOR-CANNABINOID RESULT MATTERS FOR A REASON THE STUDY WAS NOT ASKING ABOUT: after eating cannabis, THCV was “never detected in blood” at the study's analytical cut-offs, and it had cleared from oral fluid within hours. THCV present in cannabis material does not reliably reach measurable levels in the bloodstream from eating that material.
What it shows: That the THCV naturally present in cannabis was not detectable in blood after people ate a controlled dose of it.
What it does NOT show: Anything about isolated THCV taken as a product — the dose-ranging pharmacokinetic study published later gave purified THCV by mouth and measured it in plasma readily. What it does undercut is the folk claim that particular cannabis varieties deliver a meaningful THCV dose to a person: at the amounts present in the material eaten here, none reached the blood at all.
Controlled human dosing · analytical chemistry · not a THCV trialOff-registry · carries no evidence grade
How the claim travels anyway — the secondary source, named so you can check it
A refusal nobody can verify is just a louder opinion. So the review that asserts appetite suppression is named here and published beside the primary studies it summarises. We do not call it wrong; we show what it is summarising and let the two sit in the same scroll. CBG Atlas cites it as evidence for nothing.
2025
The role of tetrahydrocannabivarin (THCV) in metabolic disorders: A promising cannabinoid for diabetes and weight management
What was actually tested: A single-author narrative review of THCV in obesity and type 2 diabetes. A SECONDARY SOURCE: it reports other people's experiments and runs no study of its own.
The review states that THCV's receptor profile enables “appetite suppression, enhanced glucose regulation, and increased energy expenditure”, that animal models revealed THCV's potential to suppress appetite, and that “preliminary human trials support these findings”. ⚠ THIS PAGE PUBLISHES THE PRIMARY STUDIES THOSE SENTENCES SUMMARISE, AND THEY DO NOT CARRY THEM: the mouse study states THCV did not significantly affect food intake or body weight gain in any of its studies; the two human imaging studies each report no significant difference in subjective experience; and the human metabolic trial registered appetite as a secondary endpoint, scored it daily for thirteen weeks, and published no appetite result.
What it shows: How the appetite-suppression claim is transmitted — a secondary source stating it in summary form, without a primary human result behind it.
What it does NOT show: Any evidence for appetite suppression. A review is not a study; it inherits the strength of the papers it cites, and on this specific point those papers are null, silent or contrary. CBG Atlas does not cite this review as evidence for any claim, and neither should anything downstream of it.
Read it with this: ⚠ INCLUDED FOR TRACEABILITY, NOT FOR SUPPORT. Publishing the refusal without naming the source would make the boundary unverifiable, and an unverifiable boundary is just an assertion. Naming it lets a reader compare the summary with the primaries in the same scroll.
Narrative review · single author · secondary source · asserts more than the primaries showOff-registry · carries no evidence grade
Safety, pharmacokinetics & drug testing
The finding a THCV buyer is most likely to need.
Not the metabolism. A legal THCV product made urine drug screens read positive for THC in nearly every participant of a registered study, and a follow-up analysis documented those positives as false against a reference method. If you are subject to workplace, sporting or custody testing, that is the operative fact on this page — so it leads the section.
2023
A Two-Phase, Dose-Ranging, Placebo-Controlled Study of the Safety and Preliminary Test of Acute Effects of Oral Δ8-Tetrahydrocannabivarin in Healthy Participants
What was actually tested: A two-phase study of oral Δ8-THCV in healthy adults, registered as NCT05210634. Phase 1 was unblinded single-ascending-dose in 3 people; Phase 2 was a double-blind, randomised, within-participant crossover in 18. Participants received single acute doses of placebo and 12.5, 25, 50, 100 and 200 mg. Safety, subjective and cognitive effects were measured before dosing and up to 8 hours after. ⚠ NOTE THE ISOMER: this programme used Δ8-THCV, not the Δ9-THCV of the metabolic literature.
Most adverse events were mild (55 of 60), and the commonest was EUPHORIC MOOD. ⚠ AND THE TOP OF THE RANGE IS NOT INERT: the 100 mg and 200 mg doses showed elevations on ratings of “feel a drug effect” and “like the drug effect”, and the authors state plainly that “THC-like effects were observed at higher THCV doses”, though mild and not associated with impairment. Several doses showed a preliminary signal for improved sustained attention, but the effect was NOT dose-dependent — 12.5, 25 and 200 mg moved it while 50 and 100 mg did not, which is not the shape a real dose-response makes. ⚠ Almost every post-dose urine screen — 78 of 79 — tested POSITIVE FOR THC at 8 hours.
What it shows: That single oral doses of Δ8-THCV up to 200 mg were tolerated by healthy adults with mostly mild adverse events, and that at the top of that range people could tell they had taken something.
What it does NOT show: That THCV is non-psychoactive — the authors' own words are that THC-like effects appeared at higher doses. It also shows nothing about repeated dosing (every dose was a single acute one), nothing about people with a medical condition (participants were healthy volunteers), and nothing about efficacy for anything: this is a safety and dose-ranging study and does not test a therapeutic claim.
Read it with this: ⚠ INDUSTRY-AUTHORED. Seven of the eight authors are affiliated with Canopy Growth Corporation. ⚠ AND A CONSUMER POINT THAT MATTERS MORE THAN THE SAFETY VERDICT: a legal, non-THC product produced positive THC urine screens in nearly every participant. Anyone subject to workplace, sporting or custody drug testing should treat that as the operative finding of this study.
Dose-ranging · placebo-controlled · in people · SAFETY, not efficacyOff-registry · carries no evidence grade
2024
Pharmacokinetics of Oral Cannabinoid Δ8-Tetrahydrocannabivarin and Its Main Metabolites in Healthy Participants
What was actually tested: The pharmacokinetic arm of the same programme (NCT05210634). Twenty-one participants across six sessions received placebo and 12.5, 25, 50, 100 and 200 mg of an oral medium-chain-triglyceride formulation of Δ8-THCV; plasma from 15 of them was collected for up to 8 hours and analysed by validated two-dimensional liquid chromatography–tandem mass spectrometry.
The main metabolite detected was 11-nor-9-carboxy-Δ8-THCV. Median time to maximum concentration was roughly four to five hours for the parent compound and slightly later for the metabolite, and exposure rose about proportionally with dose. ⚠ AND THE FINDING WITH CONSUMER CONSEQUENCES: the Δ9-THCV ISOMER AND ITS CARBOXY METABOLITE WERE DETECTED IN PLASMA — despite being undetected in the formulated drug product when a third-party laboratory analysed it. The authors' own hedge is the appropriate one: the possible conversion to the Δ9 isomer “should be further studied”.
What it shows: The first published human pharmacokinetics for Δ8-THCV — that it is absorbed by mouth, peaks after several hours, and rises roughly in proportion to dose.
What it does NOT show: That the Δ8 and Δ9 isomers are interchangeable, or that Δ8-THCV converts to Δ9-THCV in the body: an isomer found in plasma but not in the product is an OBSERVATION with more than one explanation, and the authors call for further study rather than a conclusion. ⚠ The quantitative detail of how often and at what concentration Δ9-THCV appeared is in the full text and is NOT reproduced here, because this page states only what it has read in the published abstract. It also shows nothing about effect — a concentration curve is not an outcome.
Read it with this: ⚠ THE ISOMER DISTINCTION IS NOT PEDANTRY. Nearly all of THCV's metabolic literature — the mouse work, the human pilot, the unpublished Phase 2 — is about Δ9-THCV. This safety and pharmacokinetic programme is about Δ8-THCV. They are different molecules with different published pharmacology, and evidence does not transfer between them just because the abbreviation is shared.
Pharmacokinetics · in people · n = 21 enrolled, plasma from 15Off-registry · carries no evidence grade
2026
Metabolite patterns in human urine after oral administration of highly purified Δ8-THCV
What was actually tested: Urine from the same registered programme (NCT05210634), collected before dosing and pooled 0–8 hours after oral Δ8-THCV, analysed by a validated online-extraction liquid chromatography–tandem mass spectrometry method — and, in parallel, by a routine cannabinoid immunoassay of the kind used in toxicology screening.
⚠ 70 OF 80 POOLED POST-DOSE URINE SAMPLES WERE REPORTED POSITIVE BY A CANNABINOID IMMUNOASSAY TARGETING Δ9-THC-COOH — while the reference method found them NEGATIVE for Δ9-THC-COOH and positive mainly for the Δ8-THCV carboxy metabolite. The authors' conclusion is explicit: THCV's major urinary metabolites cross-react with immunoassays used routinely for toxicology testing, “resulting in false positive results for Δ9-THC exposure”.
What it shows: Directly, in people, that taking a purified THCV product can make a routine urine drug screen read positive for THC when no THC metabolite is present. Documented false positives, with a confirmatory method to prove them false.
What it does NOT show: That THCV is THC, or that anyone was exposed to THC — the reference analysis found none. It also does not establish how every immunoassay on the market behaves; one assay was tested, in one programme, on Δ8-THCV. What it establishes is that the failure mode is real and has been measured.
Read it with this: ⚠ THE PRACTICAL CONSEQUENCE IS THE POINT, AND IT IS SEVERE. A positive workplace, sporting or custody drug screen has consequences that arrive long before anyone runs a confirmatory test. This is a legal product producing a positive result for an illegal one. It is stated here plainly because a reference that publishes THCV's metabolic hopes and buries its drug-screen behaviour has chosen the seller's half of the story.
Analytical · human urine · documented immunoassay false positivesOff-registry · carries no evidence grade
It can make you fail a drug test for THC
The single most consequential practical finding on this page. In a registered human study, 78 of 79 urine screens taken 8 hours after oral THCV were positive for THC (Peters 2023, PMID 37721990) — and a follow-up analysis showed 70 of 80 pooled samples flagged positive by a routine immunoassay while the reference method found no Δ9-THC metabolite at all (Sempio 2026, PMID 42250012). If you are subject to workplace, sporting or custody drug testing, treat a THCV product as a drug-test risk regardless of what its label says about THC content.
“Non-psychoactive” does not survive the higher doses
THCV is widely marketed as non-psychoactive. In the dose-ranging safety study, the commonest adverse event was EUPHORIC MOOD, and the 100 mg and 200 mg doses produced elevated ratings of “feel a drug effect” and “like the drug effect”; the authors state that THC-like effects were observed at higher THCV doses, though mild and not associated with impairment (Peters 2023, PMID 37721990). The receptor pharmacology says the same thing in a different language: the CB1 direction reverses at higher doses.
Two different molecules share the abbreviation
Almost all of THCV's metabolic literature is about Δ9-THCV. The modern safety and pharmacokinetic programme is about Δ8-THCV (Peters 2023; Sempio 2024, PMID 39770444). They are distinct isomers with distinct published pharmacology, and a finding about one is not a finding about the other. A product label that says only “THCV” has not told you which one you are buying.
Human exposure is short, and long-term safety is uncharacterised
The longest published human exposure is thirteen weeks, in 62 people, at 5 mg twice daily (Jadoon 2016, PMID 27573936). The dose-ranging safety work gave SINGLE acute doses (Peters 2023). Nobody has published a study of anyone taking THCV for a year, no interaction study in people exists, and the 207-patient twelve-week Phase 2 that recorded adverse events, hypoglycaemia, depression and suicidality scores has never reported them.
Cannabis varieties are not a reliable THCV dose
After a controlled dose of edible cannabis, THCV was never detected in the blood of participants at the study's analytical cut-offs (Newmeyer 2017, PMID 28188235). Whatever THCV is present in plant material, eating that material did not put a measurable amount into the bloodstream — so “a high-THCV strain” is not a way to take a THCV dose.
The class this compound sits beside was withdrawn for psychiatric harm
CB1 blockers were developed as anti-obesity drugs and abandoned after depressive side effects; that history is why the unpublished Phase 2 registered the Beck Depression Inventory and the Columbia-Suicide Severity Rating Scale as outcomes, and why the human imaging studies frame THCV as a NEUTRAL antagonist rather than an inverse agonist. ⚠ THCV IS NOT THAT DRUG and no such effect has been reported for it — but the psychiatric safety data from the largest trial ever run has not been published, so the question is open rather than answered.
Educational, not medical advice
THCV has no approved use in any jurisdiction and is not a treatment for type 2 diabetes, obesity, or any other condition. Nothing here diagnoses, treats, cures or prevents any disease, and nothing here is a recommendation to take THCV or a dose at which to take it. Nobody should change or stop a prescribed diabetes medicine — metformin included — on the strength of this page; type 2 diabetes is managed by a clinician and untreated high blood glucose causes harm a supplement will not prevent. If you take prescription medicines, discuss any cannabinoid product with your prescriber or pharmacist. See safety & reporting and how we grade.
Mechanism
A direction that reverses — and one the literature disagrees about.
Receptor pharmacology, kept separate from the trial record on purpose: a binding direction and a clinical outcome are answered by different study designs, and merging them is how a laboratory result becomes a health claim. Mechanism detail is researcher-tier on this platform; switch to Researcher depth for the cited record.
In plain terms: THCV acts on the same receptors THC does, but not in the same direction — and not always in the same direction as itself. At lower amounts it gets in the way of cannabinoids that would otherwise switch the receptor on; at larger amounts, in animals, it starts switching the receptor on instead. On a second receptor, two respected papers from the same research lineage report opposite effects, and neither has been withdrawn. We publish the disagreement rather than choose a side. Switch to Researcher depth for the quoted detail.
A correction, and where it came from — read this before the cards below
One paper on this page is a graded record in the gated CBG Atlas registry: the 2008 review. Its registry keyFinding, however, records only the THC and CBD conclusions — THCV is missing from the field, even though the paper's own title names Δ9-tetrahydrocannabivarin and its abstract states four THCV-specific directions. So the directions published here are transcribed from the paper's verbatim abstract and are shown with the standing “in the registry — but graded for a different finding”, which is exactly what that label is for. They carry no evidence grade on our ladder, and none is claimed for them. A previous version of this page went the other way and reported the source as giving no THCV direction at all; it does, they are quoted in full above, and the correction is recorded at the head of lib/thcv.ts.
What the registry record carries
The gated record as the registry holds it. Note its key finding: it summarises the THC and CBD conclusions of a three-compound paper. THCV — named in the title — is not in the field.
What the paper itself says about THCV — quoted, with its dose condition
Every direction below carries the condition under which it holds, in its own column. That is deliberate and it is enforced in the data layer: the type in lib/thcv.ts makes a receptor direction without a dose condition a compile error, because “THCV is a CB1 antagonist” is a true half of a sentence whose other half reverses it.
| Target & preparation | Direction — and the source's own words | Dose condition |
|---|---|---|
| CB2 receptorsIn vitro — cells expressing the receptor. No animal and no person. | Partial agonism — THCV switches the receptor on, but not to the extent a full agonist does.“Delta9-THCV behaves as a potent CB2 receptor partial agonist in vitro.”Pertwee 2008, PMID 17828291 ⚠ THE OPPOSITE DIRECTION IS ALSO PUBLISHED, FROM THE SAME LABORATORY LINEAGE. Thomas 2005 (PMID 16205722) — whose senior author is the author of the 2008 review — concludes verbatim that “THCV behaves as a competitive CB1 and CB2 receptor antagonist”. Antagonist and partial agonist are opposite directions at the same receptor. CBG Atlas does not pick one: both are real, both are checkable, and the disagreement is published as a disagreement. | In vitro only. The source states no dose for this result, and no in-vivo dose has been shown to reproduce it in a person. ⚠ AND THE DIRECTION ITSELF IS DISPUTED — see the contested note. |
| CB1 receptors — in tissueIn vitro — tissues and cells that express CB1, challenged with cannabinoid agonists. | Antagonism — THCV blocks other cannabinoids from acting at CB1, and the strength of the block depends on which tissue and which agonist.“In contrast, it antagonizes cannabinoid receptor agonists in CB1-expressing tissues. This it does with relatively high potency and in a manner that is both tissue and ligand dependent.”Pertwee 2008, PMID 17828291 | In vitro, in the presence of an agonist. ⚠ THE SOURCE ITSELF QUALIFIES THE RESULT AS CONDITIONAL: it is tissue-dependent AND ligand-dependent, so it is not one number that travels. |
| CB1 receptors — in a living animalIn vivo — dosed animals, not isolated tissue, and not people. | ⚠ THE DIRECTION REVERSES WITH DOSE. THCV behaves as a CB1 antagonist — and at higher doses as a CB1 agonist, which is the opposite direction at the same receptor.“Delta9-THCV also interacts with CB1 receptors when administered in vivo, behaving either as a CB1 antagonist or, at higher doses, as a CB1 receptor agonist.”Pertwee 2008, PMID 17828291 | ⚠ DOSE IS THE WHOLE FINDING. The source says “at higher doses”. It does NOT say “at low dose”, and this page never writes that phrase: the low end is where antagonism was observed, but no threshold is stated, and no dose in a person has been established at all. The nearest measured dose figures are from mice, given intravenously — see Pertwee 2007 below. |
| CB1 receptors — where the dose flip was measuredIn vivo, in MICE, dosed intravenously with SYNTHETIC Δ9-THCV (O-4394) and Δ8-THCV (O-4395). | Below the flip, both compounds blocked Δ9-THC's effects. Above it, given on their own, they produced the cannabinoid-agonist-like effects they had been blocking.“By themselves, O-4395 and O-4394 induced ring immobility at 3 or 10 mg kg(-1) (i.v.) and antinociception at doses above 10 mg kg(-1) (i.v.).”Pertwee 2007, PMID 17245367 | ⚠ THE FIGURES ARE MOUSE, INTRAVENOUS, AND SYNTHETIC. Ring immobility appeared at 3 or 10 mg/kg; antinociception at doses above 10 mg/kg. THEY DO NOT CONVERT TO A HUMAN ORAL DOSE and nothing on this page treats them as guidance — they are reported because they are the primary measurement behind the phrase “at higher doses”. |
What we do NOT write, and why: “a CB1 antagonist at low dose”
It is the obvious way to summarise the reversal, and it is an inference rather than a citation. The source says “at higher doses”. It never says at low dose, it names no threshold, and the only place a threshold has been measured is in mice given THCV intravenously (Pertwee 2007). Antagonism is what was observed below the flip — but calling that “low dose” on a page a person might read next to a bottle would attach a human dose band to a mouse experiment. The sentence we can defend is the source's own, so that is the sentence we print.
The registry anchor, and the paper that contradicts it at CB2
Two verified papers give opposite CB2 directions — a partial agonist in one, an antagonist in the other — and the senior author of the earlier paper is the sole author of the later review. This is not two rival groups disagreeing; it is one research lineage reporting different things from different preparations, which is how pharmacology behaves. CBG Atlas does not resolve it. Picking the more recent, or the more flattering, or the one in the bigger journal would be an editorial judgement wearing the costume of a scientific one.
2008
The diverse CB1 and CB2 receptor pharmacology of three plant cannabinoids: delta9-tetrahydrocannabinol, cannabidiol and delta9-tetrahydrocannabivarin
What was actually tested: A single-author review of how three phytocannabinoids — Δ9-THC, CBD and Δ9-THCV — interact with cannabinoid CB1 and CB2 receptors. A SECONDARY SOURCE: it reports and synthesises other people's experiments, in cells, tissues and animals. No people, and no clinical outcome, anywhere in it.
On THCV specifically, four statements, all verbatim: “Delta9-THCV behaves as a potent CB2 receptor partial agonist in vitro.” · “In contrast, it antagonizes cannabinoid receptor agonists in CB1-expressing tissues.” · “This it does with relatively high potency and in a manner that is both tissue and ligand dependent.” · “Delta9-THCV also interacts with CB1 receptors when administered in vivo, behaving either as a CB1 antagonist or, at higher doses, as a CB1 receptor agonist.” ⚠ THE LAST ONE IS THE WHOLE MECHANISM STORY: the direction at CB1 depends on the dose, and the review says “at higher doses” — it never says “at low dose”, and neither does this page.
What it shows: That THCV's cannabinoid-receptor pharmacology has been characterised, that it antagonises CB1 agonists in CB1-expressing tissues, and that in a living animal the CB1 direction reverses with dose.
What it does NOT show: Any effect in a person, any dose for a person, and any clinical outcome. It is a review of receptor pharmacology, and receptor pharmacology does not predict what a compound does to someone. ⚠ Its CB2 statement is also CONTRADICTED by another verified paper (Thomas 2005) and is published here as contested rather than settled.
Registry status: ⚠ THIS PAPER IS IN THE GATED CBG ATLAS REGISTRY — AND ITS REGISTRY keyFinding OMITS THCV. The field records only the THC and CBD conclusions, in a three-compound paper whose title names Δ9-tetrahydrocannabivarin. The registry also supports no graded CLAIM from this record, so it carries a non-ladder method tier (“Review”) rather than a rung on our evidence ladder. Everything stated above about THCV is therefore transcribed from the paper's own verbatim abstract and carries NO evidence grade. The gap has been reported for correction through the gated pipeline; until it lands, this page reads the paper rather than the field.
Review · secondary source · CB1/CB2 pharmacology of three cannabinoidsIn the registry — but graded for a different finding
2005
Evidence that the plant cannabinoid Delta9-tetrahydrocannabivarin is a cannabinoid CB1 and CB2 receptor antagonist
What was actually tested: Radioligand displacement on mouse brain membranes and on cells expressing the human CB2 receptor; functional assays of agonist-stimulated signalling in the same membranes; and antagonism of several cannabinoid agonists in isolated mouse vas deferens. Isolated tissue and cell preparations throughout.
The paper's conclusion, verbatim: “THCV behaves as a competitive CB1 and CB2 receptor antagonist.” ⚠ AND IT IS EXPLICIT THAT THE EFFECT IS CONDITIONAL: in the vas deferens THCV antagonised several cannabinoids more potently than it antagonised THC, and it was more potent against two agonists in that tissue than in brain membranes — the authors write that “the bases of these agonist- and tissue-dependent effects remain to be established.” At high concentrations THCV also reduced contractile responses through a mechanism that was not CB1-mediated at all.
What it shows: That in isolated tissues and membranes THCV competitively blocks cannabinoid agonists at CB1 and at CB2, and that how strongly it does so depends on which tissue and which agonist.
What it does NOT show: Anything in a person, and no settled CB2 direction. ⚠ THE 2008 REVIEW REPORTS THE OPPOSITE DIRECTION AT CB2 — a partial agonist rather than an antagonist. Both statements stand in the published literature and this page presents them side by side. The binding constants in this abstract are researcher-tier figures and are deliberately not reproduced.
Read it with this: ⚠ SAME LABORATORY LINEAGE AS THE REVIEW THAT DISAGREES WITH IT. Roger Pertwee is the senior author here and the sole author of the 2008 review. The disagreement is between preparations and years, not between rival groups — which makes it more informative, not less.
In vitro · mouse brain membranes, CHO cells, isolated vas deferens · not peopleOff-registry · carries no evidence grade
2007
The psychoactive plant cannabinoid, Δ9-tetrahydrocannabinol, is antagonized by Δ8- and Δ9-tetrahydrocannabivarin in mice in vivo
What was actually tested: Synthetic Δ9-THCV (O-4394) and Δ8-THCV (O-4395) compared with THCV extracted from cannabis, first in binding and functional assays on mouse brain membranes and isolated vas deferens, then in live MICE given the compounds INTRAVENOUSLY and challenged with 3 or 10 mg/kg of Δ9-THC.
At intravenous doses of 0.1 to 3 mg/kg, both synthetic compounds attenuated THC-induced antinociception and hypothermia in mice; the Δ8 compound also attenuated THC-induced ring immobility while the Δ9 compound did not. ⚠ AND THIS IS THE MEASUREMENT BEHIND THE PHRASE “AT HIGHER DOSES”, VERBATIM: “By themselves, O-4395 and O-4394 induced ring immobility at 3 or 10 mg kg⁻¹ (i.v.) and antinociception at doses above 10 mg kg⁻¹ (i.v.).” Above that threshold the compounds produced, on their own, the cannabinoid-agonist-like effects they had been blocking below it.
What it shows: That in mice, THCV blocks THC's effects at lower intravenous doses and produces cannabinoid-agonist-like effects of its own at higher ones — the dose-dependent reversal, measured rather than asserted.
What it does NOT show: Any human dose. ⚠ THESE ARE MOUSE, INTRAVENOUS, SYNTHETIC-COMPOUND FIGURES and they do not convert to a milligram dose of an oral human product; nothing on this page treats them as guidance. They also say nothing about metabolism, appetite or weight. The binding constants in this abstract are researcher-tier figures and are deliberately not reproduced.
Preclinical · MICE, intravenous · the primary source for the dose flipOff-registry · carries no evidence grade
The registry-graded reading of the same record
CB1 receptors · direction reverses with doseReview
The registry's verified review states that Δ9-THCV antagonises cannabinoid receptor agonists in CB1-expressing tissues — with relatively high potency, and in a way that depends on both the tissue and the agonist used. Given to a live animal it interacts with CB1 as an antagonist or, at higher doses, as an agonist: the direction reverses. No threshold dose is stated, and none has been established in a person. Reported here as the review reports it, in a preclinical context, and not as a claim about any dose anyone takes.
Pertwee et al. (2008) · British Journal of PharmacologyPMID 17828291
CB2 receptors · two published directions, in conflictReview
The registry's verified review states that Δ9-THCV behaves as a potent CB2 receptor partial agonist in vitro. An earlier paper from the same laboratory lineage concluded instead that THCV is a competitive CB1 and CB2 receptor antagonist (Thomas 2005, PMID 16205722) — the opposite direction at the same receptor. Both are real, published and checkable. CBG Atlas publishes the disagreement rather than choosing a side, and asserts no CB2 direction of its own.
Pertwee et al. (2008) · British Journal of PharmacologyPMID 17828291
No potency or affinity figure appears on this page
Binding constants are researcher-tier figures on this platform and never reach a consumer surface — and here there is a second reason on top of the rule. At least three different THCV binding figures appear across the papers cited above, measured in different preparations against different reference compounds, for different receptors; and two of those papers disagree about the direction of the CB2 effect in the first place. A single number lifted out of that would look like precision and function as a fabrication.
Preclinical tracks
Rats and mice — labelled as such.
THCV has a second literature with nothing to do with metabolism: seizures, and Parkinson's disease. None of it is clinical. It is published here in full because it circulates as though it were, and because one of these results is routinely reported without the limit that sits in its own abstract.
2010
Δ9-Tetrahydrocannabivarin suppresses in vitro epileptiform and in vivo seizure activity in adult rats
What was actually tested: Rat piriform cortex brain slices with epileptiform activity induced by magnesium-free media and recorded on a multielectrode array, cannabinoid receptor binding assays in rat cortical membranes, and pentylenetetrazole-induced seizures in live adult rats. RATS THROUGHOUT.
In the slices, THCV reduced the incidence of burst complexes and the amplitude and frequency of paroxysmal depolarising shifts, both when applied to established activity and when applied beforehand. ⚠ AND THE IN-VIVO RESULT IS NARROWER THAN IT IS USUALLY REPORTED: THCV was tested across a range from 0.025 to 2.5 mg/kg, and 0.25 mg/kg — a single dose in the middle of that hundred-fold range — significantly reduced seizure incidence. NO DOSE-RESPONSE RELATIONSHIP IS REPORTED. In the binding work THCV behaved as a CB1 ligand but produced no agonist stimulation, which is the signature of a neutral antagonist rather than an activator.
What it shows: That THCV reduced epileptiform activity in rat brain tissue, and that one dose of it reduced chemically-induced seizure incidence in live rats.
What it does NOT show: Any effect in a person, and no anticonvulsant claim of any kind. ⚠ A SINGLE EFFECTIVE DOSE WITHOUT A DOSE-RESPONSE CURVE IS A WEAKER RESULT THAN A RANGE THAT MOVES TOGETHER, and this page reports it that way. There is no THCV epilepsy trial in people. Epilepsy is a serious condition managed by a specialist and nothing here is a treatment for it. The binding constants in this paper are researcher-tier figures and are deliberately not reproduced.
Preclinical · rat brain slices + rats · not peopleOff-registry · carries no evidence grade
2011
Symptom-relieving and neuroprotective effects of the phytocannabinoid Δ9-THCV in animal models of Parkinson's disease
What was actually tested: Rats with 6-hydroxydopamine lesions and mice with lipopolysaccharide lesions — two standard animal models of Parkinson's disease — given acute or chronic Δ9-THCV, alongside experiments in CB2 receptor-deficient mice. ANIMALS THROUGHOUT.
Acute THCV attenuated the motor inhibition caused by the 6-hydroxydopamine lesion, and chronic THCV reduced the loss of tyrosine hydroxylase-positive neurones in the substantia nigra — an effect the authors relate to its antioxidant properties, and one reproduced by a cannabidiol-enriched botanical extract. In the lipopolysaccharide model, where CB2 receptors were more strongly upregulated, THCV also preserved those neurones, an effect that appeared to involve CB2.
What it shows: A preclinical signal for THCV in two animal models of Parkinson's disease, on both movement and neurone survival.
What it does NOT show: Anything in a person. There is no THCV trial in Parkinson's disease, THCV is not a treatment for it, and nobody should alter Parkinson's medication on the strength of an animal model. ⚠ Note also that part of the neuroprotective effect was reproduced by a CBD extract, so it is not specific to THCV.
Preclinical · rats and mice · not peopleOff-registry · carries no evidence grade
2020
Beneficial effects of the phytocannabinoid Δ9-THCV in L-DOPA-induced dyskinesia in Parkinson's disease
What was actually tested: Δ9-THCV given intraperitoneally at 2 mg/kg for two weeks to Pitx3-mutant mice — a genetic model of dopamine deficiency — while dyskinesia was induced by repeated L-DOPA. Abnormal involuntary movements and activity were measured, along with FosB protein and a histone marker in the basal ganglia. MICE, BY INJECTION.
Given from the first L-DOPA injection, THCV delayed the appearance of dyskinetic signs and reduced their intensity, with corresponding falls in the two molecular markers. Given for three days after dyskinesia was already established, it also attenuated the signs. ⚠ THE PAPER ALSO RESTATES THE DOSE-DEPENDENCE OF THCV'S CB1 DIRECTION, IN ANIMALS: it describes THCV's CB1 antagonist profile as occurring at doses lower than 5 mg/kg. That is an animal statement about injected doses and does not convert to a human dose.
What it shows: That in a mouse model, THCV delayed and reduced the involuntary movements caused by L-DOPA.
What it does NOT show: Anything about people with Parkinson's disease or L-DOPA-induced dyskinesia. The authors themselves say further studies are required to determine the clinical significance in humans. No THCV trial in Parkinson's disease exists, and this is not a treatment.
Read it with this: ⚠ ATTRIBUTION. This paper is sometimes credited to García; its first author is ESPADAS. Concepción García is the fifth of seven authors. Recorded here because a citation attached to the wrong first author is a broken citation even when the DOI is right.
Preclinical · genetically modified MICE · not peopleOff-registry · carries no evidence grade
A single effective dose is not a dose-response — and it is reported as such
The rat seizure result is tested across a hundred-fold range, and one point in the middle of that range moved the outcome, with no dose-response relationship reported (Hill 2010). That is a weaker result than a range of doses that move together, and this page says so rather than reporting the headline and dropping the shape of the data. There is no THCV trial in people for epilepsy or for Parkinson's disease, and nothing on this page is a treatment for either.
The boundary, stated
What we deliberately do not attribute to THCV.
Six things circulate as THCV evidence and are refused here. Each refusal carries its reason and its source, so nobody has to re-derive it the next time one is pasted back in — stating a boundary once is cheaper than defending it forever, and a boundary a reader can check is worth more than one they have to take on trust.
“THCV is diet weed / an appetite suppressant”
⚠ NO HUMAN RESULT SUPPORTS THIS, AND THE RECORD IS UNUSUALLY CLEAR. Appetite was a REGISTERED SECONDARY ENDPOINT of the only metabolic trial in patients, scored daily on a 0–10 scale for thirteen weeks (NCT01217112) — and the published paper reports no appetite result at all. Two imaging studies measured subjective ratings directly and both were null (Tudge 2015, PMID 25542687; Rzepa 2016, PMID 26362774). The mouse study the claim is traced to says, verbatim, that THCV “did not significantly affect food intake or body weight gain in any of the studies” (Wargent 2013, PMID 23712280). This is not an absence of evidence; it is evidence that keeps not showing the effect.
“THCV is a proven diabetes treatment”
⚠ NO. There is no approved THCV medicine anywhere, for any indication. The human evidence is a 62-person PILOT whose registered primary endpoint (HDL) was unaffected, whose combination arms were all null, and whose glucose result comes from roughly a dozen people on THCV over thirteen weeks (Jadoon 2016, PMID 27573936). The trial that would have tested this properly — 207 patients, HbA1c as primary — completed in 2016 and was never published (NCT02053272). A blood-marker change in a pilot is not a treatment.
“THCV is non-psychoactive”
Not as stated. At 100 mg and 200 mg in a registered human study, THCV produced elevated ratings of “feel a drug effect” and “like the drug effect”, and the authors state that THC-like effects were observed at higher THCV doses; the commonest adverse event across the study was euphoric mood (Peters 2023, PMID 37721990). The receptor pharmacology agrees: the CB1 direction reverses at higher doses. “Non-intoxicating at the doses in a typical product” might be defensible; “non-psychoactive” full stop is not.
A single THCV binding figure
⚠ WE PUBLISH NO POTENCY OR AFFINITY NUMBER FOR THCV AT ALL. At least three different binding constants appear across the papers on this page, measured in different preparations, against different radioligands, for different receptors — and two of those papers disagree about the DIRECTION of the CB2 effect in the first place. A single number lifted out of that would look like precision and function as a fabrication. Binding constants are researcher-tier figures on this platform and never appear on a consumer surface.
“A high-THCV strain will give you a THCV dose”
Not supported. After a controlled dose of edible cannabis, THCV was never detected in participants' blood at the study's analytical cut-offs (Newmeyer 2017, PMID 28188235). Whatever a certificate of analysis says about THCV content in plant material, eating that material did not put a measurable amount into the bloodstream. The human studies that did measure THCV in plasma all gave a purified compound at a known milligram dose.
Δ8-THCV evidence is not Δ9-THCV evidence
The metabolic literature — the mouse studies, the human pilot, the unpublished Phase 2 — is about Δ9-THCV. The modern safety, pharmacokinetic and drug-screen work is about Δ8-THCV. Two isomers, distinct published pharmacology, one shared abbreviation. This page keeps them apart and labels every record with which one it used; nothing here transfers a finding from one isomer to the other.
THC vs THCV vs CBDV vs CBN
Four columns, because two would hide the point.
THCV against its pentyl parent, its propyl sibling, and the platform's other pillar built on a missed primary endpoint. THC, CBDV and CBN are reference chemistry here, not subjects of this page — and no column's evidence transfers to another.
| Property | THC (pentyl · reference) | THCV (this page · propyl) | CBDV (propyl) | CBN (oxidation product) |
|---|---|---|---|---|
| Side chain | Pentyl — five carbons. C₂₁H₃₀O₂ · 314.5 g/mol. | PROPYL — three carbons. C₁₉H₂₆O₂ · ≈ 286.4 g/mol. Two carbons shorter than THC. | PROPYL — three carbons. C₁₉H₂₆O₂ · ≈ 286.4 g/mol. Same formula as THCV, different skeleton. | Pentyl, on an oxidised ring. C₂₁H₂₆O₂ · ≈ 310.43 g/mol. |
| Where it comes from | CBGA → THCA synthase → THCA → heat removes CO₂ → Δ9-THC. | CBGVA → THCA synthase → THCVA → heat removes CO₂ → Δ9-THCV. ⚠ NOT MADE BY SHORTENING THC — a parallel line with its own acid, step for step. | CBGVA → CBDA synthase → CBDVA → heat removes CO₂ → CBDV. The same varin precursor, a different synthase. | Not made by the plant as an end product at all — CBN appears as Δ9-THC and THCA oxidise with air, light and time. |
| Cannabinoid-receptor direction | CB1 and CB2 partial agonist — the basis of intoxication. Reference chemistry here; THC is not a pillar on this platform. | ⚠ DEPENDS ON DOSE. Antagonises CB1 agonists in CB1-expressing tissues; in a live animal a CB1 antagonist, or at higher doses a CB1 agonist. At CB2 the literature disagrees with itself — partial agonist in one paper, antagonist in another. | Not asserted here — see the CBDV pillar for what its own record carries. | Not asserted here. Most of the cannabinoid-receptor activity after a CBN dose belongs to a metabolite rather than to CBN itself — see the CBN pillar. |
| What it was tested FOR in people | Not a pillar on this platform. | Metabolic outcomes in type 2 diabetes; blunting the effects of THC; safety and pharmacokinetics of single doses. | Seizures, HIV-associated neuropathic pain, and autism-related brain imaging. | Sleep, and essentially nothing else. Every modern CBN trial in people is a sleep trial. |
| How the trials went | — | The metabolic pilot MISSED ITS PRIMARY ENDPOINT (HDL, unaffected) and its combination arms were all null. ⚠ AND THE 207-PATIENT PHASE 2 WITH HbA1c AS PRIMARY WAS NEVER PUBLISHED. | Four published, including a Phase 2 randomised controlled trial in focal epilepsy. No primary efficacy endpoint was met. | Three that tested CBN for sleep. Both trials that pre-specified a primary endpoint MISSED IT; the third names no registration at all. |
| The claim it is sold on | — | “Diet weed” — appetite suppression and weight loss. ⚠ NO POSITIVE HUMAN APPETITE RESULT EXISTS; the trial that registered appetite as a secondary endpoint never reported it. | That it inherits CBD's clinical standing because the names and scaffolds are close. | “The sleep cannabinoid” — that a secondary or tertiary positive is the same thing as a primary endpoint. |
| The caution that travels with it | Intoxication, and everything that follows from it. | DRUG TESTS. A legal THCV product produced positive THC urine screens in nearly every participant, and documented immunoassay false positives. Plus: “non-psychoactive” does not hold at the higher tested doses. | Liver enzymes — transaminase elevations appeared in two independent CBDV studies. | Melatonin. CBN inhibited the enzyme that clears melatonin and raised melatonin exposure four-fold in mice, and the two are sold in the same capsule. |
| Approved medicine | — | None, in any jurisdiction, for any indication. | None, in any jurisdiction. | None, in any jurisdiction. |
| Does the neighbouring column transfer? | NO. | NO — and this is the column where the assumption is strongest. THCV is not “THC lite”: it is a different molecule, made from a different acid, whose receptor direction can be the OPPOSITE of THC's depending on the dose. | NO. Sharing a formula with THCV is a fact about atoms, not about evidence. | NO. Being a cannabinoid is a structural fact, not shared evidence. |
Sharing a formula is not sharing evidence
THCV and CBDV have the same molecular formula and come from the same precursor, and their evidence bases have nothing in common: CBDV has been through a Phase 2 trial in focal epilepsy, THCV through a metabolic pilot in 62 people. THCV and THC differ by two carbons and can act in opposite directions at the same receptor. Each of these four compounds has its own file, built from its own studies, and none inherits another's standing.
Research roadmap
What new evidence would change what we say about THCV?
Not an empty box, and not a request for work that has already been done. Note that the first item asks for no new science at all — only that finished science be released, which is a different and far cheaper ask than another trial.
01
Publish the trial that already exists
The single highest-value thing that could happen to THCV's evidence base requires no new participants and no new funding: 207 people completed a twelve-week Phase 2 with HbA1c as its primary endpoint in February 2016, and neither results nor a paper has ever appeared (NCT02053272). Releasing it would settle the metabolic question in either direction — and it would release twelve weeks of unreported safety, hypoglycaemia, depression and suicidality data on 207 people. Until it appears, every THCV metabolic summary in existence, including this one, is describing a subset.
02
An appetite endpoint that is actually reported
Appetite is the claim THCV is sold on and there is no positive human result for it anywhere. One trial registered it as a secondary endpoint and scored it daily for thirteen weeks without publishing the outcome; two others measured subjective ratings and found nothing. What would change this page is a trial with appetite or food intake as a PRE-SPECIFIED PRIMARY endpoint, reported whatever it shows. A null would be just as valuable as a positive, and considerably more likely.
03
Settle the CB2 direction, and find the dose flip in a person
Two verified papers give opposite CB2 directions from the same research lineage, and neither has been retracted. A modern head-to-head characterisation in one preparation would resolve it. Separately, the CB1 antagonist-to-agonist reversal has only been located in mice given THCV intravenously; nobody has established where that threshold sits for a person taking THCV by mouth — which is the difference between a dose that blocks cannabinoid signalling and one that mimics it.
04
Human data on the isomer that is actually being sold
The metabolic evidence is about Δ9-THCV; the modern safety, pharmacokinetic and drug-screen work is about Δ8-THCV. Products say “THCV”. Straightforward isomer-labelled human pharmacokinetics and repeat-dose safety for BOTH would tell a buyer which literature applies to the bottle they are holding — and would settle whether the Δ9-THCV found in plasma after a Δ8-THCV dose reflects conversion in the body, which the authors who observed it say should be studied further.
References & further reading
Where every THCV claim on this page comes from
Every paper cited above, split by standing. One is a graded record in the gated registry; the rest are real, published, checkable and ungraded here. And one entry is not a paper at all, which is exactly why it is listed.
In the gated registry — graded, and cited here for a finding the registry field omits
2008
The diverse CB1 and CB2 receptor pharmacology of three plant cannabinoids: delta9-tetrahydrocannabinol, cannabidiol and delta9-tetrahydrocannabivarin
Review · secondary source · CB1/CB2 pharmacology of three cannabinoids
THCV-specific — published, off-registry, ungraded
2016
Efficacy and Safety of Cannabidiol and Tetrahydrocannabivarin on Glycemic and Lipid Parameters in Patients With Type 2 Diabetes: A Randomized, Double-Blind, Placebo-Controlled, Parallel Group Pilot Study
Randomised, placebo-controlled · in people · PILOT · primary endpoint missed
2016
The effect of five day dosing with THCV on THC-induced cognitive, psychological and physiological effects in healthy male human volunteers: A placebo-controlled, double-blind, crossover pilot trial
Crossover pilot · in people · n = 10 · an INTERACTION study, not an efficacy trial
2015
Neural effects of cannabinoid CB1 neutral antagonist tetrahydrocannabivarin on food reward and aversion in healthy volunteers
Randomised crossover · in people · n = 20 · brain imaging, not eating
2016
The CB1 Neutral Antagonist Tetrahydrocannabivarin Reduces Default Mode Network and Increases Executive Control Network Resting State Functional Connectivity in Healthy Volunteers
Randomised crossover · in people · n = 20 · resting-state imaging
2013
The cannabinoid Δ9-tetrahydrocannabivarin (THCV) ameliorates insulin sensitivity in two mouse models of obesity
Preclinical · MICE · four dose-ranging studies · not people
2017
Cannabis Edibles: Blood and Oral Fluid Cannabinoid Pharmacokinetics and Evaluation of Oral Fluid Screening Devices for Predicting Δ9-Tetrahydrocannabinol in Blood and Oral Fluid following Cannabis Brownie Administration
Controlled human dosing · analytical chemistry · not a THCV trial
2025
The role of tetrahydrocannabivarin (THCV) in metabolic disorders: A promising cannabinoid for diabetes and weight management
Narrative review · single author · secondary source · asserts more than the primaries show
2023
A Two-Phase, Dose-Ranging, Placebo-Controlled Study of the Safety and Preliminary Test of Acute Effects of Oral Δ8-Tetrahydrocannabivarin in Healthy Participants
Dose-ranging · placebo-controlled · in people · SAFETY, not efficacy
2024
Pharmacokinetics of Oral Cannabinoid Δ8-Tetrahydrocannabivarin and Its Main Metabolites in Healthy Participants
Pharmacokinetics · in people · n = 21 enrolled, plasma from 15
2026
Metabolite patterns in human urine after oral administration of highly purified Δ8-THCV
Analytical · human urine · documented immunoassay false positives
2005
Evidence that the plant cannabinoid Delta9-tetrahydrocannabivarin is a cannabinoid CB1 and CB2 receptor antagonist
In vitro · mouse brain membranes, CHO cells, isolated vas deferens · not people
2007
The psychoactive plant cannabinoid, Δ9-tetrahydrocannabinol, is antagonized by Δ8- and Δ9-tetrahydrocannabivarin in mice in vivo
Preclinical · MICE, intravenous · the primary source for the dose flip
2010
Δ9-Tetrahydrocannabivarin suppresses in vitro epileptiform and in vivo seizure activity in adult rats
Preclinical · rat brain slices + rats · not people
2011
Symptom-relieving and neuroprotective effects of the phytocannabinoid Δ9-THCV in animal models of Parkinson's disease
Preclinical · rats and mice · not people
2020
Beneficial effects of the phytocannabinoid Δ9-THCV in L-DOPA-induced dyskinesia in Parkinson's disease
Preclinical · genetically modified MICE · not people
Not a publication — a registration with no result in the public record
NCT02053272 — “A Randomised, Double Blind, Placebo Controlled, Parallel Group, Dose Ranging Study of GWP42004 as Add on to Metformin in the Treatment of Participants With Type 2 Diabetes”. Phase 2 — randomised, double-blind, placebo-controlled, parallel-group, dose-ranging. COMPLETED. Started February 2014; primary completion February 2016; overall completion February 2016. ⚠ NONE. No paper reporting this trial has been published. A PubMed search for the compound code “GWP42004” returns ZERO records — checked 2026-08-23. It is listed here, in the references, on purpose: the reason it is missing from every other THCV reference list is that reference lists are made of publications, and this trial produced none. ClinicalTrials.gov →
For the compound this platform holds its deepest graded record on, see the CBG monograph; for THCV's sibling on the same varin precursor, see CBGV and CBDV; for the other pillar built on a set of missed primary endpoints, see CBN; and for how a claim earns a rung here, see how we grade. Peregrine Biopharma funds the platform and holds no editorial authority.
Compound monographs
The rest of the family.
Seven compounds, each with its own monograph. The line under each one is what the verified registry actually holds for it — including where that is nothing yet.
Cannabigerol (CBG)
The mother cannabinoid. CBGa decarboxylates to CBG, and both occur in the plant — it is CBGa that the three synthases branch into THC, CBD and CBC. The best-evidenced compound on this platform.
Cannabigerolic acid (CBGa)
The acid CBG comes from, and the shared substrate the three branch synthases compete for. Decarboxylates to CBG with heat.
Cannabigerovarin (CBGV)
The propyl (varin) analogue of CBG — a three-carbon side chain instead of five.
Cannabichromene (CBC)
Non-intoxicating, and one of the three branch products off the shared CBGA substrate.
Cannabinol (CBN)
Forms as THC ages and oxidises, so it accumulates in stored material rather than being made by the plant directly.
Cannabidivarin (CBDV)
The propyl analogue of CBD, and the varin with the largest published human trial record.