⚠️ Auto-extracted by Claude on 2026-05-18 — PMC OA full text (PMC12811813) used as primary source; local PDF download pending. Primary and secondary endpoint numerics extracted from PMC full text; supplementary eMethods (lifestyle co-intervention protocol detail) and full collaborative-group roster (Supplement 3) not yet cross-checked. Quantitative claims should be re-verified against the downloaded PDF on the next pass.

Lei et al. 2026 — BRAVO trial

The BRAVO trial is the largest and most methodologically rigorous randomized placebo-controlled trial of berberine for visceral and hepatic adiposity conducted to date. It enrolled diabetes-free adults with obesity and metabolic dysfunction-associated steatotic liver disease (MASLD) across 11 Chinese hospitals and delivered 6 months of treatment at the standard clinical dose (1 g/day). The trial was negative on both co-primary endpoints — berberine produced no significant reduction in visceral adipose tissue (VAT) area or liver fat content compared to placebo. Crucially, both arms also received individualized lifestyle counseling, which drove weight loss in both groups; this co-intervention makes it impossible to attribute body-composition changes to berberine alone.

Secondary lipid endpoints were positive: berberine reduced LDL-cholesterol (−7.72 mg/dL, p=0.008) and apolipoprotein B (−3.42 mg/dL, p=0.02), effects mechanistically consistent with the post-transcriptional LDLR mRNA stabilization mechanism documented by Kong et al. 2004. These lipid effects survived placebo control, providing the clearest human-RCT evidence that berberine’s cholesterol-lowering activity is mechanism-driven and distinct from any body-composition effect.

Trial registration and design

ParameterDetail
Trial nameBRAVO (Berberine in Adiposity)
RegistrationNCT05647915
PhasePhase 3 (implied by design; n=337 multicenter RCT)
DesignMulticenter, double-blind, randomized, placebo-controlled
Sites11 hospitals in China
Enrollment windowJuly–December 2023
Follow-up6 months
AnalysisIntention-to-treat (full analysis set, n=337) with multiple imputation for missing data; per-protocol sets for primary endpoints (reduced to 283 for VAT, 269 for liver fat after exclusions)

Randomization: 1:1 using minimized randomization via a central online system, stratified by age, sex, prediabetes status, abnormal liver function, and lipid-lowering medication use.

Blinding: Investigators, participants, data managers, CT technicians, and radiologists were all blinded to allocation.

Intervention arm: Berberine hydrochloride 0.5 g orally twice daily (1 g/day total) for 6 months.

Control arm: Matching placebo capsules 0.5 g twice daily for 6 months.

Co-intervention (both arms): Tailored lifestyle interventions including individualized physical activity guidance and dietary counseling per ACC/AHA and AACE obesity management standards. The full protocol is in eMethods 3 of Supplement 1 (not independently extracted here). needs-replication — the extent and fidelity of lifestyle delivery across 11 sites was not separately characterized in the main paper; intensity may have varied.

Adherence: Berberine arm 90.3% (SD 14.7%); placebo arm 90.7% (SD 17.4%). Excellent and balanced.

Population

  • n=337 randomized (169 berberine, 168 placebo); 337 completed 6-month follow-up
  • Treatment discontinuation: 5 (3.0%) berberine vs 6 (3.6%) placebo (p=0.75, not significant)
  • Mean age: 41.8 years (SD 10.6) — a notably young-ish MASLD cohort
  • Sex: 221 male (65.6%), 116 female (34.4%)
  • Mean BMI: 31.8 kg/m² (SD 3.5)
  • Mean waist circumference: 104.1 cm (SD 10.2)
  • Mean liver fat (MRI-PDFF): 17.5% (SD 7.8%)
  • All participants were diabetes-free (key population difference from most prior berberine metabolic trials, which enrolled type 2 diabetes populations)

Primary endpoints — NEGATIVE

Both co-primary endpoints were prespecified as co-primary, each adjudicated at a 2.5% one-sided significance threshold (97.5% CI reported).

VAT area (relative percentage change from baseline)

GroupMean change
Placebo−2.0% (SD 14.4%)
Berberine−0.6% (SD 16.7%)
Between-group difference+1.38% (97.5% CI −2.43% to +5.18%; p=0.42)

Interpretation: both arms reduced VAT slightly (placebo arm more than berberine arm), attributable to the lifestyle co-intervention received by both. Berberine added no incremental benefit. NEGATIVE.

Liver fat content (absolute percentage-point change from baseline)

GroupMean change
Placebo−1.1% (SD 5.1%)
Berberine+0.1% (SD 6.2%)
Between-group difference+0.87 percentage points (97.5% CI −0.39 to +2.13; p=0.12)

Interpretation: placebo arm modestly reduced liver fat (likely lifestyle effect); berberine arm showed a trivial non-significant increase relative to placebo. NEGATIVE. The direction numerically favors placebo on both primary endpoints — not a close null, but also not statistically significant.

Secondary endpoints

EndpointDifference (berberine − placebo)95% CIpInterpretation
LDL-cholesterol (mg/dL)−7.72−13.13 to −1.930.008Significant; berberine superior
Apolipoprotein B (mg/dL)−3.42−6.33 to −0.510.02Significant; berberine superior
hs-CRP (mg/dL)−0.072−0.140 to −0.0040.04Significant; berberine superior
TriglyceridesNot significantNSNo difference
Body weight (kg)−1.9 kg (placebo) vs −1.8 kg (berberine)0.90No difference
BMI−0.7 vs −0.60.90No difference
Waist circumference (cm)−2.8 vs −2.70.91No difference
HbA1cNot significantNSNo difference
Fasting glucoseNot significantNSNo difference
Insulin resistanceNot significantNSNo difference
Blood pressureNot significantNSNo difference
Liver fibrosis markersNot significantNSNo difference

hs-CRP subgroup: Authors report that participants with elevated baseline hs-CRP showed greater anti-inflammatory benefit from berberine, consistent with berberine’s known NF-κB suppression activity.

Adverse events

CategoryBerberine (n=169)Placebo (n=168)p
Serious AEs6 (3.6%)2 (1.2%)0.28
Non-serious AEs of interest17 (10.1%)13 (7.7%)0.57
Liver impairment1 (0.6%)0>0.99
Renal impairment2 (1.2%)1 (0.6%)>0.99
Constipation1 (0.6%)0>0.99
GI reactions (other)5 (3.0%)3 (1.8%)0.72
Hypoglycemia8 (4.7%)9 (5.4%)0.81

Serious AE count was nominally higher in the berberine arm (6 vs 2), but the difference was not statistically significant (p=0.28) and the trial was not powered to detect safety differences at this event rate. The overall adverse-event profile was benign and balanced. Safety conclusion (authors): berberine was well tolerated with no excess risk of adverse events.

Significance — what this trial tells us

1. Berberine is not a meaningful weight-loss or body-composition drug under controlled conditions

The most clinically important finding is the null body-composition result against an active lifestyle co-intervention. Prior open-label and weakly-controlled berberine weight-loss trials often showed reductions in BMI and waist circumference. BRAVO demonstrates that when both arms receive structured lifestyle support — the standard of care for MASLD — berberine adds nothing to body-composition outcomes. This is the clearest human RCT evidence against berberine as an adiposity treatment.

Contrast with berberine’s efficacy in type 2 diabetes: Yin et al. 2008 (doi:10.1016/j.metabol.2007.11.007) showed substantial glycemic separation (HbA1c reduction) in T2D patients, a population with impaired baseline glucose metabolism where berberine’s AMPK activation has a mechanistically clear substrate. The BRAVO population (diabetes-free) lacks this substrate. The diabetes-free vs T2D distinction matters for expected effect size. contradictory-evidence — the glycemic null result in BRAVO (diabetes-free) is not contradicted by Yin 2008 (T2D); these are different populations.

2. Berberine’s lipid-lowering effect is mechanism-driven and survives placebo control

The LDL-C (−7.72 mg/dL) and ApoB (−3.42 mg/dL) reductions are small in absolute terms but statistically robust and mechanistically interpretable. The LDLR mRNA stabilization mechanism described by Kong et al. 1 operates through post-transcriptional ERK-mediated 3’UTR stabilization, independent of dietary fat intake or weight loss. The fact that this effect emerged in BRAVO despite the neutral body-composition result confirms it is mechanism-driven, not secondary to weight change.

This is the human-RCT validation that was previously missing for Kong 2004’s mechanistic claim. needs-replication — one RCT; independent replication at this dose would strengthen confidence.

3. MASLD-specific design gap

Berberine has been hypothesized to reduce hepatic fat via AMPK-mediated lipogenesis suppression (via ACC/FASN downregulation) and via bile acid modulation through the gut microbiome. BRAVO directly tests these in the MASLD population and finds no clinically meaningful liver fat effect over 6 months at 1 g/day. Whether higher doses, longer duration, or combination with a non-lifestyle-standard-of-care comparator would reveal an effect remains open. dose-response-unclear

DimensionStatus
Pathway conserved in humans?yes — LDLR mRNA stabilization via ERK; AMPK activation both documented in human cell lines
Phenotype conserved in humans?partial — lipid-lowering: yes (confirmed here); adiposity/liver fat: no meaningful effect vs controlled comparator
Replicated in humans?LDL-C lowering: yes (multiple prior RCTs + BRAVO); adiposity: negative (BRAVO is first adequately powered trial with lifestyle co-intervention)

Limitations (per authors and per study design)

  1. Lifestyle co-intervention in both arms — this is the dominant limitation for the adiposity interpretation. Structured dietary counseling + physical activity guidance in both arms produced comparable weight loss (~1.8–1.9 kg) in both groups, leaving no statistical headroom for berberine to separate. A placebo arm without any lifestyle intervention would have been ethically impermissible but would have yielded a cleaner adiposity contrast.
  2. 6-month duration — MASLD regression may require longer treatment windows, particularly for fibrosis. The trial was not designed to assess liver fibrosis endpoints (e.g., liver stiffness, biopsy-based NAS score).
  3. Chinese hospital population — multicenter but geographically homogeneous; external validity to Western MASLD populations is uncertain given dietary and microbiome differences (berberine’s bile-acid mechanism is microbiome-mediated and microbiome composition differs markedly across populations).
  4. 1 g/day dose only — no dose-response arm. Whether 2–3 g/day or higher could produce adiposity effects is not addressed.
  5. Young-ish cohort — mean age 41.8 years; most aging-wiki-relevant MASLD populations skew older. Results may not fully extrapolate to the 65+ population where metabolic reserve is lower and liver regeneration capacity differs.
  6. No mechanistic substudies — no liver biopsies, no LDLR expression measurements, no AMPK activity assays. The mechanistic attribution for positive secondary endpoints (LDL-C, ApoB) is inferred from Kong 2004, not directly demonstrated in this trial.

long-term-unknown — longest duration of this trial is 6 months; no data on MASLD fibrosis, cardiovascular events, or hepatocellular carcinoma progression.

Cross-references

PageRelationship
berberinePrimary compound studied; compound page (stub pending — no molecules/compounds/berberine.md yet)
kong-2004-berberine-ldlr-mechanismOrigin paper for LDLR mRNA stabilization mechanism; BRAVO confirms the lipid-lowering effect in a controlled RCT
ldlrDirect downstream target; mRNA stabilization drives LDL-C reduction
ampk-activatorsBerberine’s AMPK-activation mechanism; BRAVO does not test AMPK activation directly
deregulated-nutrient-sensingHallmark context — AMPK is a key nutrient sensor; impaired AMPK activity in metabolic disease
chronic-inflammationhs-CRP reduction in berberine arm — anti-inflammatory secondary outcome

Footnotes

Footnotes

  1. kong-2004-berberine-ldlr-mechanism · n=32 (human arm) · open-label pilot · model: hypercholesterolemic adults; HepG2 cells; hyperlipidemic hamsters · doi:10.1038/nm1135 · origin paper for LDLR mRNA stabilization mechanism