Cannabis and the Liver: What the Hepatotoxicity Data Actually Shows

Search this subject and you will find two headlines that seem to contradict each other. One says cannabis is hard on the liver. One says cannabis protects it.

Both are describing real studies. They are also describing different substances, taken in different ways, at doses that differ by a factor of a hundred. Getting the distinction right is the whole subject.

This guide covers what the evidence says about smoked cannabis, what it says about oral CBD, the drug combination that matters far more than dose, and what the fatty liver headlines are actually built on.

Two substances, two answers

Smoked cannabis has no documented liver injury signal. High-dose oral CBD does. The federal government's own liver injury database scores them differently on exactly that basis, and the single biggest risk factor is not the cannabinoid at all. It is what else the patient is taking.

What the record says about cannabis itself

The National Institutes of Health maintains a database that grades every substance on how likely it is to cause clinically apparent liver injury. Cannabis is graded near the bottom.

Its assessment is blunt: cannabis does not appear to cause acute liver injury or to exacerbate preexisting liver disease. Marijuana use has not been linked to serum enzyme elevations during therapy or to instances of clinically apparent liver injury with jaundice, and while rare case reports exist, the database's verdict on them is that none were convincing or well documented.

The isolated-molecule evidence agrees. Dronabinol, pharmaceutical THC, produced aminotransferase elevations in 6 percent of treated patients against 4.3 percent of controls, a difference with no clinical meaning, and carries the same unlikely-cause rating. Nabilone, the other synthetic THC medicine, scores identically.

Two prescription THC drugs and decades of population use, and the honest summary is that there is no human evidence base for THC hepatotoxicity.

Where the real signal is

CBD is a different molecule with a different profile, and the same database grades it accordingly. Prescription cannabidiol carries a likelihood score of E* (unproven but suspected rare cause of clinically apparent liver injury, particularly with high doses). In the epilepsy trials, elevations above three times the upper limit of normal occurred in 13 percent of cannabidiol-treated patients compared to 1 percent on placebo.

Those trial doses are enormous by consumer standards, running 10 to 25 milligrams per kilogram of body weight. For an adult that is roughly 700 to 1,750 milligrams a day, where a typical retail CBD oil delivers 25 to 50.

The same source notes that the lower doses found typically in over-the-counter CBD products are generally well tolerated without evidence of liver injury. Which was the comfortable version of this story until last year.

The trial that moved the line

In 2025 the FDA's own research division ran a randomized placebo-controlled trial in healthy adults at a dose it described as consumer-range.

Participants took roughly 400 milligrams a day for 28 days. Alanine aminotransferase rose above three times the upper limit of normal in 8 participants (5.6%; 95% CI, 1.8%-9.3%) on CBD against none on placebo, and seven participants met withdrawal criteria for potential drug-induced liver injury.

Two details keep this in proportion. The elevations took three weeks or more to appear, peaked one to two days after dosing stopped, and returned to baseline within one to two weeks. Nobody met the standard criteria for serious liver injury.

Set that against a meta-analysis of CBD trials which found that high doses and concurrent antiepileptic drugs were the risk factors, and that no cases were reported in adults using cannabidiol doses <300 mg/day.

Put the two together and a workable line appears somewhere between 300 and 400 milligrams a day. Below it, nothing shows up in the literature. Above it, something does.

The unexamined day is a wasted opportunity. Reflect on what you did, what you learned, and how you can improve.

John Dewey

The factor that matters more than dose

Here is the number that should change behavior, and it has nothing to do with how much CBD anyone takes. In the epilepsy trials, the rate of significant liver enzyme elevation was 30 percent in patients taking both valproate and clobazam, 21 percent on valproate alone, 4 percent on clobazam alone, and 3 percent on neither.

Three percent to thirty percent, driven entirely by co-medication. That is a tenfold swing, and it is larger than any dose effect in the same dataset.

The mechanism is only half understood. CBD triples blood levels of clobazam's active metabolite by inhibiting the enzyme that clears it, which is textbook pharmacology. Valproate is stranger: CBD does not raise valproate levels at all, yet the combination produces the highest enzyme elevations on record. Nobody has explained that one.

CBD is a moderate inhibitor of CYP2C19 and a weak inhibitor of CYP1A2, and inhibits several other cytochrome enzymes. The practical list of drugs to raise with a doctor before adding CBD includes warfarin, tacrolimus, clopidogrel, theophylline and the anticonvulsants above.

The fatty liver headlines, examined

Now the studies point the other way, because they are widely quoted and thinly built. A large analysis of 2014 US hospital discharge records, published in PLOS ONE, reported that cannabis use was associated with a lower prevalence of fatty liver disease, with an adjusted odds ratio of 0.82 for all users and 0.49 for dependent users, apparently dose-responsive.

The authors state their own limitation clearly, noting that the cross-sectional design inhibits their ability to draw direct causal relationships. There is a deeper problem. This is a billing database, so it measures coded cannabis use among hospitalized people, and coding for cannabis use is notoriously incomplete.

The decisive objection is that the same database and the same year produce the opposite result on a different outcome. A separate analysis of the same data, in the World Journal of Hepatology, found cannabis users hospitalized with fatty liver disease had a higher prevalence of ascites, 4.5 percent against 3.6 percent.

One dataset, two papers, two directions. Neither supports a causal claim in either direction, and anyone citing the protective finding without the counterpart is quoting half a coin toss.

The hepatitis C question, resolved

This one has a cleaner answer, and it went the way of the null. A 2005 cross-sectional study found daily cannabis use associated with faster fibrosis progression in hepatitis C, and the finding stuck in clinical folklore for years.

Then prospective cohorts followed people over time. A study in Clinical Infectious Diseases tracking 690 co-infected patients found no evidence that marijuana smoking accelerates progression to significant liver fibrosis, and a later cohort following 575 women for a median of eleven years found cumulative THC use unrelated to fibrosis progression while alcohol use in the same cohort predicted it clearly.

The likeliest explanation for the original result is reverse causation. People whose liver disease is worsening feel worse, and some of them increase cannabis use to manage it. A single snapshot cannot tell those apart. Following people for a decade can.

The one real risk for smokers

There is a genuine hazard here, and it applies to a specific group. Smoked plant material carries fungal spores, and for patients whose immune systems are suppressed, including some with advanced liver disease or post-transplant, inhaled molds are a documented danger. Researchers writing in Supportive Care in Cancer concluded that systematic sterilization of medicinal cannabis can eliminate the risk of fatal opportunistic infections among patients at risk.

For a healthy adult this is close to irrelevant. For an immunosuppressed one, combustion is the wrong format regardless of anything else in this article.

Mold is also the one part of this a grower can design against. Dense, resin-heavy flowers rot from the inside when humidity climbs, so structure and flowering time matter as much as any curing technique. Thin Mint Frosting finishes in 56 to 63 days at 100 to 140 cm indoors, Thin Mint crossed with Banana Frosting at 30% THC with 550 grams a square meter, and a shorter flowering window is fewer weeks for humidity to find a way in.

The grower's contribution to this

Where a seed company genuinely touches this subject is contamination, not chemistry.

The fungal risk above is a storage and curing problem before it is anything else. Flower dried too fast in humid air, jarred before it is ready, or left sealed without burping develops mold in exactly the conditions a home grower controls completely. Two habits do most of the work:

• Dry slowly, jar late. Aim for a week or more of drying at moderate humidity, then burp the jars daily for the first fortnight. Rushing the dry is what forces moisture back out of the stems later.
• Grow what suits your room. London Pound Cake Auto is our highest-yielding autoflower at 28% THC, 75 to 80 days from seed at 90 to 130 cm, creamy vanilla over sweet berries, and it finishes fast enough to dodge a damp autumn entirely.

Nobody in a commercial supply chain can watch your jars for you, and nobody needs to.

Where the line actually falls

The NIH liver injury database rates cannabis an unlikely cause of clinically apparent liver injury and finds the existing case reports unconvincing. Two prescription THC drugs carry the same rating. There is no human evidence base for THC hepatotoxicity.

Oral CBD is different. Epilepsy trials at 10 to 25 mg/kg produced liver enzyme elevations in 13 percent against 1 percent on placebo, and an FDA trial at roughly 400 mg a day in healthy adults found 5.6 percent against zero. A meta-analysis found no cases below 300 mg a day, putting the practical line between the two.

Co-medication matters more than dose: enzyme elevation ran 30 percent in patients on both valproate and clobazam against 3 percent on neither. The fatty liver studies claiming cannabis is protective come from a hospital billing database that produces the opposite result on a different outcome, and the hepatitis C alarm from 2005 did not survive two prospective cohorts.

Barney's Farm has been developing premium cannabis genetics since the 1980s, with over 40 Cannabis Cup wins. Explore our full seed catalog and find strains bred for every climate and skill level.