Milk thistle is the oldest and best-selling liver supplement in the world, and the single most useful thing to know about it is the shape of its evidence rather than any one result. Pooled across every trial, it looks like it reduces liver-related death. Pooled across only the well-conducted trials, that reduction disappears. The largest, best-designed randomized trial ever run on it — at doses well above what the bottles recommend — found nothing at all. And the one use that genuinely commands clinical respect, as an antidote in death-cap mushroom poisoning, involves an intravenous drug that is not the capsule and is not approved in the United States.
What milk thistle actually is
The plant is Silybum marianum. What is extracted from its seeds is silymarin, which is not a single molecule but a complex mixture of roughly six major flavonolignans plus minor polyphenols.[3] The most-studied fraction is silibinin (also called silybin), and that distinction carries most of the confusion in this category: the impressive clinical stories usually belong to purified, intravenous silibinin, while the product on the shelf is an oral silymarin extract of variable composition. Standardization varies between manufacturers, which is one reason the trial literature is so hard to pool — the trials are not all testing the same thing.
The best-designed trial: null, at above-label doses
The strongest test milk thistle has ever received was the multicenter, double-blind, placebo-controlled SyNCH trial, published in JAMA in 2012. It enrolled 154 people with chronic hepatitis C whose interferon-based therapy had failed, and gave them silymarin at 420 mg or 700 mg three times daily — deliberately higher than customary doses — or placebo, for 24 weeks.[1] The primary outcome was met by exactly two participants in each of the three groups: 3.8% on placebo, 4.0% on 420 mg, 3.8% on 700 mg, with P at or above .99. The mean decline in ALT did not differ significantly across groups (P = .75), and there were no significant differences in HCV RNA levels or quality-of-life measures either. The adverse-event profile matched placebo, so this is not a safety story; it is an efficacy story. When a supplement is given at more than its usual dose, in a properly powered and blinded trial, and moves nothing, that result deserves more weight than the dozens of small positive studies that preceded it.
The Cochrane pattern: the benefit lives in the weak trials
The most instructive analysis is Cochrane’s review of milk thistle in alcoholic and hepatitis B or C liver disease, covering 18 randomized trials in 1,088 patients.[2] Milk thistle had no significant effect on all-cause mortality (RR 0.78, 95% CI 0.53 to 1.15), on complications of liver disease (RR 0.95, 95% CI 0.83 to 1.09), or on liver histology. Liver-related mortality was significantly reduced when all trials were pooled (RR 0.50, 95% CI 0.29 to 0.88) — and that is the number the marketing quotes. But restricted to the high-quality trials, the same outcome was no longer significant (RR 0.57, 95% CI 0.28 to 1.19). The reviewers noted that only 28.6% of the trials reported high methodological quality.
This is the single most transferable lesson on the page, and it recurs across this whole market: an effect that shrinks or vanishes as trial quality rises is usually a property of the trials, not of the compound. We have watched the same thing happen to lactoferrin, where every increase in trial size cost it effect, and it is why a pooled estimate should never be read without asking what it was pooled from.
NASH: a missed primary endpoint with a real fibrosis signal
The most interesting positive result is also an honest failure. A randomized, double-blind, placebo-controlled trial in Kuala Lumpur gave silymarin 700 mg three times daily, or placebo, to 99 adults with biopsy-proven nonalcoholic steatohepatitis and a NAFLD activity score of 4 or more, for 48 weeks, with repeat biopsies at the end.[3] It missed its primary efficacy endpoint: 32.7% of the silymarin group versus 26.0% of the placebo group, P = .467. But significantly more silymarin patients showed a reduction in fibrosis of at least one point on histology (22.4% versus 6.0%, P = .023), and by liver stiffness measurement (24.2% versus 2.3%, P = .002).
Read that carefully in both directions. A fibrosis improvement confirmed on paired biopsy is not a trivial finding, and it is more than most liver supplements have ever produced. But it is a secondary endpoint in a trial that failed its primary one, in 99 patients at a single center, and secondary endpoints in failed trials are where false positives live. It is a genuine reason to run a larger trial. It is not a reason to claim silymarin treats NASH.
Blood sugar: a striking number the authors themselves distrust
A systematic review and meta-analysis of five randomized trials in 270 patients with type 2 diabetes reported that silymarin significantly reduced fasting blood glucose by 26.86 mg/dL (95% CI −35.42 to −18.30) and HbA1c by 1.07 (95% CI −1.73 to −0.40), with no effect on the lipid profile.[4] An HbA1c drop of that size would be a remarkable finding for a herbal extract — comparable to a real glucose-lowering drug — which is exactly why the authors’ own conclusion matters more than the point estimate. They wrote that, being aware of the low quality of the available evidence and the elevated heterogeneity of the studies, no recommendation can be made. A large effect measured across five small, heterogeneous, low-quality trials is a hypothesis. The same caution applies here as to berberine, where a genuine glucose effect is routinely inflated into a prescription-drug substitute.
The mushroom antidote: real, and not what you are buying
Milk thistle’s reputation rests heavily on one dramatic use: treatment of poisoning by amatoxin-containing mushrooms, chiefly Amanita phalloides, the death cap. The claim has a real basis, and it is also the most consistently misused fact in the category, for two reasons. The first is the form. The treatment is intravenous silibinin, a purified flavonolignan given in hospital — not an oral silymarin capsule — and in the United States it has never been available as an FDA-approved pharmaceutical preparation.[5] The second is the strength of the evidence. A 2022 review in Clinical Toxicology notes that although retrospective analyses favor silibinin or penicillin over routine care, there is no quality randomized trial, and that mortality from Amanita ingestion has remained near 10% across four decades of its use.
For the oral form specifically, the most recent evidence is sobering. An analysis of 255 patients with hepatotoxic mushroom poisoning at a single poison center, of whom 33 received silymarin, found that after adjustment for confounders silymarin was not associated with reduced mortality, nor with reduced acute liver failure or hospital length of stay — though the incidence of acute kidney injury was lower (adjusted risk difference −13.9%, 95% CI −23.6% to −4.1%, P = 0.005).[6]That is retrospective, single-center data and the authors say so. It is still the opposite of the confident story the supplement aisle tells about this plant.
The honest bottom line
Milk thistle is safe, cheap and genuinely old, and it is not a fraud: the fibrosis result in NASH is real, and silibinin’s place in the treatment of amatoxin poisoning is real. What is not supported is the thing it is actually sold for. There is no trial showing that a milk thistle capsule detoxifies a healthy liver, protects one from alcohol, or improves any hard outcome in ordinary use — the best-designed trial found nothing at doses above the label, and the pooled mortality benefit evaporates once you look only at the trials that were well run. Treat it as a compound with one hospital use it cannot deliver in a capsule, one interesting secondary endpoint that needs replication, and a marketing story far ahead of both.