asxan.ai Agents Hub

Milk Thistle

Evidence Fact Sheet

Silymarin

Milk thistle (Silymarin / Silybum marianum) is a flavonolignan botanical studied for antioxidant, anti-inflammatory and hepatocyte-membrane-stabilizing mechanisms, mainly in liver-disease and type-2-diabetes populations. Meta-analyses show liver-enzyme reductions but low certainty; the strongest NASH histology RCT missed its primary endpoint. Legal supplement ingredient; no authorized health claim.

Also known as: Milk Thistle · Silymarin · Silybin · Silibinin · Silybum marianum

Overview

Milk thistle is the seed extract of Silybum marianum; its active fraction, silymarin, is a complex of flavonolignans (silybin/silibinin being the principal one). Mechanistic research describes free-radical scavenging and Nrf2/glutathione up-regulation, suppression of NF-kB / TNF-alpha / COX-2 signaling, hepatocyte membrane stabilization, and anti-fibrotic (TGF-beta) effects, with insulin-signaling modulation reported mostly in type-2-diabetes populations. RCT-validated research doses cluster around 420-600 mg/day silymarin in 2-3 divided oral doses (the highest trial dose was 840 mg/day); oral bioavailability is low and phospholipid/Phytosome complexes are used to improve it, and there is no formal Upper Limit. It is a legal, well-established dietary-supplement ingredient (FDA DSHEA structure/function claims permitted with disclaimer; EU traditional-herbal registration; ANVISA herbal/food-supplement framework; China health-food raw material and OTC liver drug), but EFSA, FDA, ANVISA and China SAMR have authorized 0 single-ingredient health claims. Almost all efficacy evidence comes from liver-disease or diabetic patients, not healthy people; this page reports research findings, not medical advice.

Mechanism of Action

Antioxidant: free-radical scavenging, inhibition of lipid peroxidation, and up-regulation of the Nrf2/GSH pathway (mechanistic research context) · Anti-inflammatory: suppression of NF-kB / TNF-alpha / COX-2 signaling (preclinical and biomarker research) · Hepatocyte membrane stabilization: flavonolignans intercalate into membrane phospholipids, with membrane-stabilizing and protein-synthesis effects observed in research · Anti-fibrotic: inhibition of TGF-beta and Kupffer-cell activation (preclinical models) · Insulin-signaling modulation: effects on insulin sensitivity/resistance observed in research (mostly in T2DM populations)

Body systems: Liver & Detoxification · Endocrine & Metabolic · Cellular Renewal

Evidence-Based Benefits

Each benefit below is anchored to a specific PubMed-indexed study. Effect sizes, sample sizes, and p-values are reported as published; no values are inferred. Honest negatives and null results are kept alongside the positive findings, and disease-research populations are described as such — Milk Thistle is not characterized as a treatment for any disease.

Liver Enzymes in MASLD (Fatty Liver) — Cochrane Meta-Analysis

Meta-analysis supported
  • MD -7.21 U/LALT · 6 trials, 409 ppl
  • very low-certaintyGRADE evidence

In the 2025 Cochrane systematic review of adults with metabolic dysfunction-associated steatotic liver disease, silymarin monotherapy modestly lowered ALT versus placebo, but the certainty of evidence was very low and the authors concluded the overall benefits and harms remain unclear. Silymarin-complex formulations showed no significant ALT effect.

Reported effect: ALT: MD -7.21 U/L, 95% CI -10.62 to -3.80; 6 trials, 409 participants; very low-certainty evidence (silymarin monotherapy). Silymarin complex ALT: MD -1.10 U/L, 95% CI -8.12 to 5.92.

“ALT reduction: "MD -7.21 U/L, 95% CI -10.62 to -3.80; 6 trials, 409 participants; very low-certainty evidence"; the authors concluded that "the benefits and harms of silymarin in adults with MASLD are unclear."”

Source: PMID 40552569 · Wang 2025 · Cochrane Database Syst Rev

Liver Enzymes & Lipids in NAFLD/NASH — Meta-Analysis

Meta-analysis supported
  • SMD -12.39ALT · 95% CI -19.69,-5.08
  • SMD -10.97AST · 95% CI -15.51,-6.43

A 2024 meta-analysis of 26 RCTs in NAFLD/NASH found silymarin reduced both ALT and AST and improved several metabolic markers, including total cholesterol, LDL-C, triglycerides and fasting insulin, plus higher odds of histological improvement. Effect sizes were large but heterogeneity across enzyme outcomes was high.

Reported effect: ALT SMD -12.39 (95% CI -19.69 to -5.08); AST SMD -10.97 (95% CI -15.51 to -6.43); histological improvement OR 3.25 (95% CI 1.80 to 5.87).

“decreased levels of ALT (SMD[95%CI]=-12.39[-19.69, -5.08]); decreased levels of AST (SMD[95% CI]=-10.97[-15.51, -6.43]); Histological improvement: OR=3.25 (95% CI: 1.80 to 5.87).”

Source: PMID 38579127 · Li 2024 · Ann Hepatol

NASH Histology (NAFLD Activity Score) — Primary Endpoint of Largest Dedicated RCT

Null / no benefit RCT supported
  • 32.7% vs 26.0%NAS response · P=.467
  • 22.4% vs 6.0%fibrosis improve · P=.023

In a randomized trial of biopsy-proven nonalcoholic steatohepatitis, the primary endpoint — a 30%+ reduction in NAFLD Activity Score — did NOT differ significantly between silymarin and placebo. A secondary fibrosis endpoint did improve significantly on both histology and liver-stiffness measures, but this is an honest negative on the trial's pre-specified primary outcome.

Reported effect: Primary outcome (>=30% NAS reduction): 32.7% silymarin vs 26.0% placebo, P=.467 (not significant). Secondary fibrosis improvement: histology 22.4% vs 6.0%, P=.023.

“"The percentage of patients achieving the primary efficacy outcome did not differ significantly between the groups (32.7% in the silymarin group vs 26.0% in the placebo group; P = .467)."”

Source: PMID 28419855 · Wah Kheong 2017 · Clin Gastroenterol Hepatol

Glycemic Control in Type 2 Diabetes — Meta-Analysis

Meta-analysis supported
  • -26.86 mg/dLfasting glucose
  • -1.07HbA1c · 95% CI -1.73,-0.40
  • 5 RCTs, 270participants

A 2016 meta-analysis of type-2-diabetes RCTs found silymarin significantly reduced fasting blood glucose and HbA1c versus control. The authors flagged the low quality and high heterogeneity of the underlying trials, stated no recommendation could be made, and noted silymarin had no effect on lipid profile in this analysis.

Reported effect: Fasting blood glucose -26.86 mg/dL (95% CI -35.42 to -18.30); HbA1c -1.07 (95% CI -1.73 to -0.40); 5 RCTs, 270 participants.

“"routine silymarin administration determines a significant reduction in fasting blood glucose levels (-26.86 mg/dL; 95% CI -35.42-18.30)" and "HbA1c levels (-1.07; 95% CI -1.73-0.40)"; "no recommendation can be made and further studies are needed."”

Source: PMID 27340676 · Voroneanu 2016 · J Diabetes Res

Insulin Resistance & Sensitivity — Meta-Analysis

Meta-analysis supported
  • WMD -2.29HOMA-IR · p=0.047
  • WMD -2.56fasting insulin · p=0.862 (NS)
  • 6 RCTs, 673participants

A 2025 meta-analysis found silymarin produced a statistically significant but modest improvement in HOMA-IR, with no significant effect on fasting insulin and no meaningful change in the QUICKI insulin-sensitivity index. The result is mixed: insulin-resistance index improved while direct insulin measures did not.

Reported effect: HOMA-IR WMD -2.29 (95% CI -4.55 to -0.03, p=0.047); fasting insulin WMD -2.56 (95% CI -7.60 to 2.48, p=0.862, not significant); 6 RCTs, 673 participants.

“"Meta-analysis showed that Silybum marianum significantly improved HOMA-IR (WMD = -2.29, 95 % CI: -4.55 to -0.03, p = 0.047)"; it "had no effect on FI (WMD = -2.56, 95 % CI: -7.60 to 2.48, p = 0.862)."”

Source: PMID 39855603 · Yin 2025 · Diabetes Res Clin Pract

Liver Enzymes by Cause of Liver Injury — Meta-Analysis (Subgroup Negatives)

Null / no benefit Meta-analysis supported
  • SMD -0.912ALT overall · p=0.000
  • SMD -0.670AST overall · p=0.000

A 2025 meta-analysis of 55 trials found silymarin significantly lowered ALT and AST overall, but the subgroup analysis was an honest negative: enzymes fell in NAFLD and viral hepatitis yet showed NO significant hepatoprotective effect in drug-induced or alcohol-related liver injury. The authors suggested benefit may depend on dose and duration.

Reported effect: Overall ALT SMD -0.912 (95% CI -1.177 to -0.646, p=0.000); AST SMD -0.670 (95% CI -0.931 to -0.408, p=0.000); no significant effect in drug-induced or alcohol-related liver injury subgroups.

“"AST and ALT levels significantly decreased in patients with NAFLD and viral hepatitis, while it had no significant hepatoprotective effects in patients with drugs induced liver injury and alcohol-related liver disease."”

Source: PMID 40221681 · Shahsavari 2025 · BMC Complement Med Ther

Nephroprotection (Kidney) — Meta-Analysis

Meta-analysis supported
  • Hedges g -1.23creatinine · p=0.0024
  • 7 studiesclinical trials
  • I² = 93.40%high heterogeneity

A 2025 meta-analysis reported a significant overall reduction in serum creatinine with silymarin, driven by drug-induced acute kidney injury; there was no significant effect in chronic kidney disease and the 280 mg/day dose was ineffective. Heterogeneity was very high, so the signal is subgroup-dependent rather than general.

Reported effect: Serum creatinine reduction Hedges' g -1.23 (95% CI -2.02 to -0.43, p=0.0024, I²=93.40%); significant in drug-induced AKI (p=0.003) but not in CKD (p=0.3065).

“"meta-analysis was performed on 7 studies and showed a significant effect of silymarin on the reduction of serum creatinine levels (Hedges' g, -1.23; 95% CI, -2.02 to -0.43; p = 0.0024; I² = 93.40%)"; CKD subgroup "no significant effect (p = 0.3065)."”

Source: PMID 40886393 · Frounchi 2025 · Biochemistry (Mosc)

Dosage (research context · not a recommendation)

RCT-validated doses 420-600 mg/day silymarin (2-3 divided oral doses); the highest trial dose was 840 mg/day (Wah Kheong 2017 used 700 mg x3/day). No formal Upper Limit. Oral bioavailability is low (~20-50%); phospholipid/Phytosome complexes can improve it. Educational reference, not a dosing recommendation.

Regulatory Status · 4 Markets

US · FDA
Legal, well-established dietary-supplement ingredient. DSHEA structure/function claims (e.g. supports liver health / antioxidant support) permitted with the mandatory FDA disclaimer; disease claims prohibited. No FDA-authorized health claim.
EU · EFSA
Marketable in several member states (traditional-herbal registration / HMPC recognition of traditional use); EFSA has authorized 0 health claims (the MILTE ITALIA 2010 Art.13.5 lactation claim was rejected).
CN · China
China SAMR: legal health-food raw material; silybin also approved as OTC liver drug. Marketable as supplement ingredient; no single-ingredient authorized health function asserted.
BR · ANVISA
Usable under the ANVISA traditional-herbal-medicine / food-supplement framework; no IN 28/2018 Anexo V authorized functional claim.

Safety

Good safety profile (confirmed by Cochrane and multiple meta-analyses; liver-patient trials dosed up to 840 mg/day, GI upset <5%). Caution in people allergic to Asteraceae (ragweed, daisy) plants. May affect some drug metabolism via CYP3A4/UGT, so people on concomitant medication should consult a professional. Almost all efficacy evidence comes from liver-disease or T2DM patient populations; direct evidence in healthy people is limited. This page describes scientific evidence, not medical advice; people under treatment for liver disease or diabetes should not self-substitute it for medical care.

Goals: longevity-stack · inflammation-relief

Lifestyles: intermittent-fasting · senior-60-plus

References

PubMed-indexed citations anchoring the benefit findings above. Effect sizes are reported as published.

  1. PMID 40552569 · Wang 2025 · Cochrane Database Syst Rev — Liver Enzymes in MASLD (Fatty Liver) — Cochrane Meta-Analysis
  2. PMID 38579127 · Li 2024 · Ann Hepatol — Liver Enzymes & Lipids in NAFLD/NASH — Meta-Analysis
  3. PMID 28419855 · Wah Kheong 2017 · Clin Gastroenterol Hepatol — NASH Histology (NAFLD Activity Score) — Primary Endpoint of Largest Dedicated RCT
  4. PMID 27340676 · Voroneanu 2016 · J Diabetes Res — Glycemic Control in Type 2 Diabetes — Meta-Analysis
  5. PMID 39855603 · Yin 2025 · Diabetes Res Clin Pract — Insulin Resistance & Sensitivity — Meta-Analysis
  6. PMID 40221681 · Shahsavari 2025 · BMC Complement Med Ther — Liver Enzymes by Cause of Liver Injury — Meta-Analysis (Subgroup Negatives)
  7. PMID 40886393 · Frounchi 2025 · Biochemistry (Mosc) — Nephroprotection (Kidney) — Meta-Analysis

Frequently Asked Questions

1. Does milk thistle (silymarin) actually lower liver enzymes?

Meta-analyses do report reductions in ALT and AST in studied liver-disease populations — for example the 2024 NAFLD/NASH meta-analysis reported ALT SMD -12.39 and AST SMD -10.97. But the 2025 Cochrane review of MASLD rated the same kind of ALT signal (MD -7.21 U/L) as very low-certainty and concluded the benefits and harms remain unclear. So the direction is consistent across studies, but the strength of evidence is weak. This is research evidence in patients, not a treatment claim.

2. Is there a study where milk thistle did NOT work?

Yes. The largest dedicated randomized trial in biopsy-proven nonalcoholic steatohepatitis (Wah Kheong 2017) missed its primary endpoint: a 30%+ reduction in NAFLD Activity Score occurred in 32.7% of the silymarin group vs 26.0% of placebo (P=.467, not significant). Separately, the 2025 55-trial meta-analysis found no significant enzyme benefit in drug-induced or alcohol-related liver injury. These honest negatives matter as much as the positive signals.

3. What about blood sugar and diabetes?

In type-2-diabetes RCTs pooled in a 2016 meta-analysis, silymarin significantly reduced fasting blood glucose (-26.86 mg/dL) and HbA1c (-1.07), though the authors stressed low study quality and made no recommendation. A 2025 meta-analysis found a modest significant improvement in HOMA-IR (WMD -2.29, p=0.047) but no significant effect on fasting insulin. Evidence is concentrated in diabetic patients, not healthy people, and this is not dosing guidance.

4. What dose was used in research and is it a supplement?

RCT-validated research doses cluster around 420-600 mg/day silymarin in 2-3 divided oral doses, with the highest trial dose at 840 mg/day; oral bioavailability is low, which is why phospholipid/Phytosome complexes are used, and there is no formal Upper Limit. Milk thistle is a legal, well-established supplement ingredient (FDA DSHEA, EU traditional-herbal, ANVISA, China health-food), but no regulator has authorized a single-ingredient health claim. This is an educational reference, not a dosing recommendation.

Last evidence review: 2026-06-13

← Evidence library

Ask Agent Axor