⚠️ Auto-extracted by Claude on 2026-05-18. Citation metadata confirmed via PubMed efetch (PMID 31926918); n and primary endpoint numbers confirmed from structured PubMed abstract. Full PDF is paywalled (Lancet Gastroenterol Hepatol, closed OA); quantitative claims sourced from abstract-level data. Verify exact table-level subgroup data and secondary endpoint details against primary source before relying on them. no-fulltext-access
Berberine versus placebo for the prevention of recurrence of colorectal adenoma
Chen YX, Gao QY, Zou TH et al. (Fang JY, senior author) · Lancet Gastroenterol Hepatol 5(3):267–275 · 2020 · DOI: 10.1016/S2468-1253(19)30409-1
Citation impact (DOI lookup, 2026-05-18): 186 citations · FWCI = 19.39 · citation_percentile = 100
TL;DR
A multicentre, double-blind RCT in 891 Chinese adults with prior colorectal adenoma (post-polypectomy) found that berberine hydrochloride 300 mg twice daily for up to 2 years reduced adenoma recurrence by 23% relative risk reduction (RR 0.77, 95% CI 0.66–0.91; p=0.001) compared with placebo. Advanced adenoma recurrence was also reduced (~50% relative, precise CI wider). Berberine was well tolerated; constipation was the main adverse event. The study is the largest high-quality RCT of berberine for a clinically meaningful cancer-prevention endpoint in humans. See tan-2025-berberine-adenoma-6yr-followup for the same cohort’s 6-year extended follow-up.
Background and rationale
Colorectal cancer (CRC) is one of the leading causes of cancer mortality globally. Colonic adenomatous polyps are established precursor lesions; recurrence after endoscopic polypectomy is common (~40–50% at 2 years), and identifying effective chemoprevention agents is a clinical priority. berberine is an isoquinoline alkaloid used in traditional Chinese medicine with documented lipid-lowering and anti-inflammatory properties (see kong-2004-berberine-ldlr-mechanism for the LDL-receptor-upregulation mechanism). Mechanistic work and smaller trials had suggested anti-proliferative and pro-apoptotic effects in colorectal cell lines, motivating this Phase 3-scale RCT.
Study design
| Feature | Detail |
|---|---|
| Design | Multicentre, double-blind, randomised, placebo-controlled |
| Sites | Multiple centres in China (Fudan / Shanghai Jiao Tong-affiliated) |
| Participants | Adults with prior colorectal adenoma confirmed by colonoscopy and pathology within 6 months of enrolment |
| Randomisation | 1:1 berberine vs. placebo; method not confirmed from abstract unsourced |
| Blinding | Double-blind (participants and outcome assessors) |
| Intervention | Berberine hydrochloride 0.3 g (300 mg) orally twice daily |
| Comparator | Matched placebo twice daily |
| Duration | Up to 2 years |
| Primary endpoint | Adenoma recurrence at follow-up colonoscopy |
| Key secondary endpoint | Advanced adenoma recurrence (adenoma with high-grade dysplasia, ≥25 mm, or villous/tubulovillous histology) |
| Safety | Adverse event monitoring throughout; no serious adverse events reported |
Results
Primary outcome
In the full-analysis set (n=891; 429 berberine, 462 placebo):
- Adenoma recurrence: 36% (155/429) berberine vs. 47% (216/462) placebo
- Unadjusted RR = 0.77 (95% CI 0.66–0.91; p=0.001) 1
- Absolute risk reduction: ~11 percentage points
- Number needed to treat (NNT): approximately 9 (derived; not confirmed from primary text) unsourced
Advanced adenoma
Advanced adenoma recurrence was reduced by approximately 50% (relative) in the berberine arm. The precise RR, CI, and p-value for this endpoint are not confirmed from abstract-level data — verify against Table 2 or equivalent in the primary PDF. no-fulltext-access
Safety
- Constipation was the most common adverse event: 6 participants in the berberine group vs. 1 in placebo group
- No serious adverse events reported in either arm
- No reports of hepatotoxicity or other major organ-level toxicity at this dose and duration
| Dimension | Status |
|---|---|
| Pathway conserved in humans? | yes — human RCT; no extrapolation required |
| Phenotype conserved in humans? | yes — adenoma recurrence is a hard clinical endpoint |
| Replicated in humans? | in-progress — same cohort 6-year follow-up published 2025; independent replication pending |
Context and limitations
Strengths:
- Largest berberine adenoma-prevention RCT published as of 2020; double-blind; hard clinical endpoint
- n=891 well-powered for the primary outcome
- Dose (300 mg BID) is pharmacologically relevant and matches commonly used berberine supplement doses
- 2-year duration adequate for adenoma recurrence as a clinical endpoint
- CRC incidence monitored (no cases detected in either arm over the trial period — too short and underpowered for CRC as primary endpoint)
Limitations:
- Single ethnic cohort — all participants from Chinese centres; colorectal adenoma biology, dietary background, gut microbiome, and CYP2D6/CYP3A4 pharmacogenetics may differ in non-Asian populations needs-replication
- Adenoma recurrence is a surrogate endpoint for CRC, not CRC itself — long-term CRC prevention benefit not established by this trial alone
- Mechanistic pathway not established in this trial — berberine’s anti-adenoma effect may be mediated by ampk activation, anti-inflammatory activity, LDL-R modulation (see kong-2004-berberine-ldlr-mechanism), or gut-microbiome remodelling; this trial does not disambiguate no-mechanism
- Constipation signal (6 vs. 1) warrants monitoring in larger/longer trials, especially in older participants or those on anticholinergic co-medications
- Full safety table, secondary endpoint table, and subgroup data (sex, age, adenoma burden) not confirmed from abstract — require primary PDF verification no-fulltext-access
Significance and wiki role
This study is the primary human RCT anchor for berberine’s potential anti-cancer activity in colorectal settings. It is among the most-cited berberine RCTs in aging-adjacent literature (186 citations, FWCI=19.4, citation_percentile=100 per OpenAlex). The Fang JY group (Fudan/Shanghai Jiao Tong) published a 6-year extended follow-up in 2025 — see tan-2025-berberine-adenoma-6yr-followup — which should be read in conjunction with this page for the long-term dataset.
This study contributes evidence to:
- berberine — compound page (canonical home for mechanism, PK, full trial roster)
- cancer — phenotype page; CRC adenoma prevention arm
Cross-references
| Page | Relationship |
|---|---|
| berberine | Compound page; this study is the primary RCT anchor for berberine’s anti-cancer human evidence |
| cancer | Phenotype; CRC adenoma is the clinical outcome tested |
| tan-2025-berberine-adenoma-6yr-followup | 6-year extended follow-up of the same cohort; must be read with this page |
| kong-2004-berberine-ldlr-mechanism | Mechanistic companion; berberine’s LDL-R upregulation mechanism |
| ampk | Putative mechanistic mediator; AMPK activation is one proposed route for berberine’s anti-proliferative effects |
Limitations and gaps
no-fulltext-access — Lancet Gastroenterol Hepatol is paywalled; full PDF not downloaded as of 2026-05-18. Subgroup data, secondary endpoint tables, and randomisation details not confirmed from primary text. Verifier should obtain institutional access or author request.
needs-replication — All enrolled participants from Chinese centres. Generalisability to non-Asian populations is untested. An independent Western replication or pooled analysis is needed before broad chemoprevention recommendations.
no-mechanism — The trial does not report mechanistic biomarkers. Berberine’s anti-adenoma mechanism (AMPK, gut microbiome, inflammation, lipid-receptor pathway) is not established in this dataset.
long-term-unknown — 2-year trial duration is adequate for adenoma recurrence but insufficient to assess CRC incidence, long-term safety, or durability of benefit after cessation. See tan-2025-berberine-adenoma-6yr-followup for extended data.
Footnotes
Footnotes
-
chen-2020-berberine-adenoma-rct (this page) · n=891 (429 berberine, 462 placebo) · rct · p=0.001 · model: human adults, post-polypectomy colorectal adenoma · abstract-level data confirmed via PubMed efetch PMID 31926918; full PDF not verified no-fulltext-access ↩