Berberine
Berberine is an AMPK-activating plant alkaloid with a metformin-class effect in type 2 diabetes: HbA1c −0.71% across 27 RCTs (Lan 2015 meta-analysis, n=2,569), LDL −25 mg/dL (Koppen 2017). Hard-contraindicated in pregnancy (kernicterus) and clinically significant CYP3A4/2D6 inhibitor.
Berberine scored 6.8 / 10 (👍 Worth trying) on the BioHarmony scale as a Substance → Botanical Extract (non-adaptogenic).
What It Is
Berberine is an isoquinoline alkaloid extracted from Berberis aristata (Indian barberry), Coptis chinensis (goldthread), Hydrastis canadensis (goldenseal), and Phellodendron amurense bark. Traditional Chinese and Ayurvedic medicine used it for centuries as an antimicrobial for dysentery and skin infections. Modern clinical interest exploded after the Yin 2008 Metabolism head-to-head trial showed near-equivalence to metformin in type 2 diabetes, and TikTok's 2023-2024 "nature's Ozempic" wave turned a metabolic alkaloid into a mass-market weight loss supplement.
Type: Botanical isoquinoline alkaloid (OTC supplement).
Current status: Nick uses occasionally for targeted glycemic support. Quality-certified HCl only (NSF or USP). Cycles on and off rather than daily continuous.
Terminology
- AMPK: AMP-activated protein kinase. Cellular energy sensor triggered by low ATP.
- PCSK9: Proprotein convertase subtilisin/kexin type 9. Protein that degrades LDL receptors; inhibiting it lowers LDL.
- PTP1B: Protein tyrosine phosphatase 1B. Inhibits insulin receptor signaling; blocking PTP1B restores insulin sensitivity.
- NPC1L1: Niemann-Pick C1-Like 1 protein. Intestinal cholesterol transporter, same target as ezetimibe.
- GLP-1: Glucagon-like peptide-1. Gut hormone that stimulates insulin and reduces appetite.
- HbA1c: Glycated hemoglobin. 3-month average blood glucose marker.
- HOMA-IR: Homeostatic Model Assessment of Insulin Resistance. Fasting glucose times fasting insulin divided by 405.
- IRS-1: Insulin Receptor Substrate 1. Key signaling protein downstream of the insulin receptor.
- PEPCK: Phosphoenolpyruvate carboxykinase. Rate-limiting gluconeogenesis enzyme.
- G6Pase: Glucose-6-phosphatase. Final gluconeogenesis enzyme.
- GLUT4: Glucose transporter type 4. Insulin-regulated glucose transporter.
- NAFLD: Non-alcoholic fatty liver disease.
- PCOS: Polycystic ovary syndrome.
- SIBO: Small intestinal bacterial overgrowth.
- DHB: Dihydroberberine. Reduced form of berberine with higher bioavailability.
- AUC: Area under the curve. Total drug exposure over time on a pharmacokinetic plot.
- Cmax: Maximum plasma concentration after a dose.
- CYP3A4 / CYP2D6 / CYP2C9: Cytochrome P450 enzyme subtypes that metabolize roughly 75% of pharmaceuticals.
- P-gp: P-glycoprotein. Drug efflux pump in the gut lining and blood-brain barrier.
- OATP: Organic anion-transporting polypeptide. Hepatic drug uptake transporters.
- NOAEL: No Observed Adverse Effect Level. Highest tested dose with no observed harm in toxicology studies.
- DILI: Drug-Induced Liver Injury.
- FAERS: FDA Adverse Event Reporting System.
- EFSA: European Food Safety Authority. EU equivalent to FDA for food and supplements.
- NDA Panel: EFSA Panel on Nutrition, Novel Foods and Food Allergens.
- RCT: Randomized controlled trial.
- SMD: Standardized mean difference. Effect size metric in meta-analyses.
- WMD: Weighted mean difference. Same units as original measurement.
- CGM: Continuous glucose monitor.
- T2D: Type 2 diabetes.
Dosing & Protocols
Dosing information is summarized from published research and community reports. This is not a prescribing guide. Consult a healthcare provider before starting any protocol.
View 1 route and 5 protocols
Routes & Forms
| Route | Form | Clinical Range | Community Range |
|---|---|---|---|
| Oral | 1.0-1.5 g/day berberine HCl (500 mg TID with meals) | 500 mg once daily up to 3.0 g/day TID |
Protocols
Standard T2D / prediabetes Clinical
- Dose
- 500 mg HCl three times daily with meals (1.5 g/day)
- Frequency
- TID
- Duration
- 8-12 weeks, recheck labs, continue if indicated
Validated by Yin 2008 (PMID 18442638) and Lan 2015 meta (PMID 25498346). Start 500 mg once daily for 1 week to assess GI tolerance before titrating to TID.
PCOS insulin resistance Clinical
- Dose
- 500 mg HCl three times daily with meals
- Frequency
- TID
- Duration
- Minimum 12 weeks
Wei 2012 meta shows comparable effect to metformin on androgens and ovulation. Combine with myo-inositol 4 g/day + D-chiro 100 mg. Cease immediately upon positive pregnancy test.
SIBO-D short course Community
- Dose
- 400-500 mg HCl three times daily
- Frequency
- TID
- Duration
- 4 weeks, not for chronic use
Often stacked with oregano oil or alongside rifaximin. Berberine is broadly antimicrobial; prolonged use disrupts beneficial taxa.
Dihydroberberine alternative Anecdotal
- Dose
- 100-200 mg twice daily with meals
- Frequency
- BID
- Duration
- Same as standard HCl protocol
~5x plasma AUC versus HCl (Wang 2017, PMID 28836457). No long-term efficacy RCTs. Treat as off-label improvement.
Biohacker cycling Anecdotal
- Dose
- 1.5 g/day HCl
- Frequency
- TID
- Duration
- 8 weeks on, 4 weeks off
No RCT validates cycling. Community rationale: microbiome preservation, CYP450 normalization. Continuous use is widely practiced without acute harm.
Use-Case Specific Dosing
| Use Case | Dose | Notes |
|---|---|---|
How this score is calculated →
Upside (2.27 / 5.00)
| Dimension | Weight | Score | Visual | Weighted |
|---|---|---|---|---|
| Efficacy | 25% | 3.5 | 0.875 | |
| Breadth of Benefits | 15% | 3.8 | 0.570 | |
| Evidence Quality | 25% | 3.5 | 0.875 | |
| Speed of Onset | 10% | 3.0 | 0.300 | |
| Durability | 10% | 2.0 | 0.200 | |
| Bioindividuality Upside | 15% | 3.0 | 0.450 | |
| Total | 3.270 |
Upside Rationale
Efficacy (3.5/5.0): Berberine reduces HbA1c by 0.71% and fasting glucose by 1.21 mmol/L across 27 RCTs (Lan 2015 meta-analysis, n=2,569), matching first-line oral diabetes agents in indicated populations. Yin 2008 head-to-head (n=84) showed HbA1c −2.0% versus metformin −1.8%. Koppen 2017 lipid meta: LDL −25 mg/dL, triglycerides −44 mg/dL. For metabolically healthy users chasing weight loss, pooled effect collapses to 1-2 kg. The 3.5 score averages a bimodal response: ~4.2 in T2D/PCOS/NAFLD, ~2.0 in healthy adults.
Breadth (3.8/5.0): Berberine has validated effects across metabolic, cardiovascular, hepatic, reproductive, and gut domains (5+ distinct indication clusters with RCT support). Wei 2012 PCOS meta (535 women) showed ovulation and androgen improvements comparable to metformin. Yan 2015 NAFLD RCT (n=184, PMID 26002205) showed ALT −36% and steatosis grade improved in 63% versus 31% control. Documented antimicrobial activity against H. pylori, SIBO methanogens, C. difficile, and Candida. Early colorectal adenoma signals (Chen 2020, Lancet Gastroenterol Hepatol) are interesting but narrow.
Evidence (3.5/5.0): 50+ RCTs and 3+ meta-analyses with consistent directional findings, but 90%+ of trials are Chinese single-center with n<100 and <12 week duration (Koppen 2017 explicitly flagged high I2 and publication bias). v0.5 evidence integrity penalty is −0.5 default (no industry burial, no failed replication), not stacked. No Cochrane review exists. Fewer than 5 independent Western-led replications. Long-term (>12 month) data is sparse outside the Chen 2020 colorectal cohort.
Speed (3.0/5.0): Fasting glucose drops within 1-2 weeks on 1.5 g/day; HbA1c requires 8-12 weeks per the 3-month erythrocyte turnover window (Lan 2015 methodology). Lipid effects plateau at 4-8 weeks per Koppen 2017. Antimicrobial effects on H. pylori and SIBO are measurable in 24-72 hours. Middle-tier speed: not fast like caffeine or GLP-1 injectables, not slow like resistance training.
Durability (2.0/5.0): Benefits persist only while dosed. Effects fade within 4-8 weeks of discontinuation per community CGM reports and the Yin 2008 run-out observations. No evidence of metabolic reprogramming that outlasts cessation. Berberine corrects dysfunction while present; it does not resolve the underlying insulin resistance permanently. No disease-modifying claim survives scrutiny.
Bioindividuality (3.0/5.0): Responders skew heavily toward dysregulated metabolism: T2D, prediabetes, PCOS, NAFLD, metabolic syndrome. The responder-versus-non-responder spread is roughly 4x on HbA1c delta (Lan 2015 subgroup analyses). Healthy users with normal fasting glucose see minimal shift. GI tolerance varies 5-10x across individuals, partly mediated by gut microbiome composition that converts HCl to dihydroberberine.
Downside (1.16 / 5.00)
| Dimension | Weight | Score | Visual | Weighted |
|---|---|---|---|---|
| Safety Risk | 30% | 2.5 | 0.750 | |
| Side Effect Profile | 15% | 2.8 | 0.420 | |
| Financial Cost | 5% | 1.4 | 0.070 | |
| Time/Effort Burden | 5% | 1.8 | 0.090 | |
| Opportunity Cost | 5% | 1.5 | 0.075 | |
| Dependency / Withdrawal | 15% | 1.0 | 0.150 | |
| Reversibility | 25% | 1.2 | 0.300 | |
| Total | 1.855 | |||
| Harm subtotal × 1.4 | 2.268 | |||
| Opportunity subtotal × 1.0 | 0.235 | |||
| Combined downside | 2.503 | |||
| Baseline offset (constant) | −1.340 | |||
| Effective downside penalty | 1.163 |
Downside Rationale
Safety (2.5/5.0 raw, 3.5 effective after 1.4x harm amplifier): Three documented harm vectors justify the amplifier. First, berberine displaces bilirubin from albumin ~10x more potently than phenylbutazone (PMID 8513024), with documented kernicterus cases in Chinese neonates: NIH MotherToBaby rates it unsafe in pregnancy and breastfeeding. Second, Guo 2012 in-human pharmacokinetics at 300 mg TID showed CYP3A4 midazolam AUC +40%, CYP2D6 dextromethorphan ratio +9x, CYP2C9 losartan ratio doubled, and cyclosporine AUC +34.5% in transplant recipients. Third, the EFSA January 2026 draft opinion cannot establish a safe intake level, citing in-vitro genotoxicity and rodent carcinogenicity.
Side Effects (2.8/5.0): Gastrointestinal adverse events occur in 10-35% of users across 27 trials (LiverTox meta, n=2,569). Cramping, diarrhea, constipation, bitter taste, nausea. Effects are dose-dependent and mitigated by TID-with-meals dosing. Rare idiosyncratic hepatotoxicity case reports exist, typically resolving on discontinuation. David Sinclair publicly switched away from berberine by 2025 citing GI. Separate from but not stacked onto the safety floor.
Cost (1.4/5.0): Standard berberine HCl costs $15-25 per month at 1.5 g/day from NSF or USP certified vendors on Amazon or iHerb (accessible channel pricing per v0.5). Dihydroberberine (Glucovantage) runs $40-60/month. Cheap overall. The real cost is quality control: independent analysis found average berberine content at 75% of label claim (range 33-100%), and gram-per-day intake of Chinese bulk supply raises heavy metal exposure concerns at California Prop 65 thresholds.
Effort (1.8/5.0): 500 mg three times per day with meals requires 3 daily adherence events. Timing with food is not optional: fasted dosing blunts the glycemic hit and worsens GI symptoms. Forgetting doses collapses effect because berberine's plasma half-life is roughly 4 hours. Pill burden is real but not extreme.
Opportunity (1.5/5.0): Minimal direct opportunity cost. Downstream, berberine's CYP3A4 inhibition can corrupt pharmacokinetics of more important interventions (statins, cyclosporine, tacrolimus, DOACs). In metabolically healthy populations, the opportunity cost of berberine versus more durable interventions (strength training, sleep, fasting in the longevity stack) is real but not massive. v0.5 opportunity multiplier is 1.0x, not stacked onto safety.
Dependency (1.0/5.0): Feng 2015 pharmacokinetic work pegs oral berberine plasma half-life at ~9 hours, and Yin 2008 run-out observations show the glucose-lowering effect decaying over 3-5 days without compulsion, craving, or withdrawal syndrome. v0.5 separates functional reliance from addictive dependence: berberine fails both floors. Unlike caffeine, there is no adenosine receptor downregulation and no rebound fatigue. T2D-indicated users have a functional need to continue treatment similar to metformin reliance (roughly equivalent reversible dependence on exogenous glucose support), but biohackers can cycle 8 weeks on, 4 weeks off without any stack lock-in, per Zhang 2008 metabolic endpoint RCT. Cessation is pharmacologically and behaviorally clean.
Reversibility (1.2/5.0): Feng 2015 documents oral berberine clearing plasma with a ~9 hour half-life, and Dong 2012 meta-analysis follow-up data show HbA1c returning to baseline within roughly 12 weeks of discontinuation (the 3-month erythrocyte turnover window). No irreversible tissue adaptations are documented. Gut microbiome shifts recover over several weeks once antimicrobial pressure is removed, per microbiota composition studies. Hepatotoxicity case reports resolve on discontinuation (LiverTox Likelihood Score E). Unlike long-term statins (post-market adaptations in cholesterol synthesis machinery) or SSRIs (persistent receptor sensitivity changes, PSSD in a minority), berberine is fully reversible. Genotoxicity and carcinogenicity signals, if real at human exposures, would not reverse: those weight into safety, not reversibility.
Verdict
✅ Best for: Diagnosed type 2 diabetes or prediabetes patients who cannot tolerate or access metformin. PCOS with insulin resistance under clinician oversight (cease immediately upon positive pregnancy test). NAFLD with elevated ALT seeking an adjunctive option. Short-course (4-week) SIBO-D protocols. Statin-intolerant patients needing targeted LDL reduction via the PCSK9 and NPC1L1 pathways.
❌ Avoid if: You are pregnant, breastfeeding, trying to conceive, or caring for a neonate: the kernicterus risk is a hard contraindication, not cautionary labeling. You take CYP3A4 substrates (most statins except rosuvastatin and pravastatin, macrolides, many SSRIs, tacrolimus, cyclosporine, warfarin, DOACs, or complex polypharmacy) without pharmacist clearance. You are metabolically healthy and chasing weight loss or generic longevity: the evidence does not support the hype, and the safety profile does not justify it. EU residents should pause purchases until the EFSA opinion is finalized after the May 2026 consultation closes.
Use-case dependence is the headline. Berberine's score reflects a wide efficacy spread between T2D-indicated populations (subrating 8.2) and metabolically healthy users chasing hype (subrating 3.2). A single aggregate rating averages a bimodal response.
Use Case Breakdown
The overall BioHarmony score reflects the intervention's primary evidence profile. These subratings are independent assessments per use case.
| Use Case | Score | Summary |
|---|---|---|
| ✅ Blood Sugar / Glycemic Control | 8.2 | Yin 2008 non-inferior to metformin (HbA1c −2.0%); Lan 2015 meta HbA1c −0.71% across 27 RCTs |
| ✅ Metabolic Health | 8.0 | Strongest evidence domain. HbA1c, fasting glucose, lipids, NAFLD all improve in dysregulated populations |
| 💪 Hormonal / Endocrine | 7.6 | Wei 2012 PCOS meta: testosterone and LH/FSH improvements comparable to metformin |
| 💪 Liver / Detoxification | 7.5 | Yan 2015 NAFLD RCT: ALT −36%, steatosis grade improved in 63% vs 31% control |
| 👍 Cardiovascular | 6.8 | LDL −25 mg/dL, TG −44 mg/dL per Koppen 2017 meta; PCSK9 mechanism distinct from statins |
| 👍 Fertility (Female) | 6.0 | PCOS ovulation restoration via insulin sensitization (Wei 2012, An 2014) |
| 👍 Body Composition / Fat Loss | 5.8 | Peng 2023 meta: BMI −0.45 kg/m2, waist −1.23 cm. Modest, dysregulated populations only |
| ⚖️ Geriatric / Aging Population | 5.5 | Metabolic benefit applies; CYP interactions more salient with polypharmacy |
| ⚖️ Gut Health / Microbiome | 5.4 | Antimicrobial against H. pylori and SIBO methanogens; chronic use disrupts beneficial taxa |
| ○ Anti-Inflammatory | 4.5 | CRP −0.3 to −0.8 mg/L in elevated-baseline RCTs; NF-kB mechanism |
| ○ Mitochondrial | 4.2 | Complex I inhibition is mechanism of action, not net mito benefit; AMPK downstream |
| ○ Skin / Beauty | 4.2 | Topical antimicrobial history; thin oral evidence for acne |
| ○ Healthspan | 4.0 | Metabolic disease reduction extrapolates to healthspan; no direct data |
| ○ Autophagy | 4.0 | AMPK-driven autophagy induction in preclinical models; no direct human measurement |
| ○ Immune Function | 3.6 | NF-kB and cytokine suppression in vitro; thin direct human immunity evidence |
| ○ Antioxidant / Oxidative Stress | 3.5 | Nrf2 activation in preclinical models; modest clinical translation |
| ○ Longevity / Lifespan | 3.4 | AMPK and mTOR suppression theoretical longevity pathway; animal lifespan data; no human data |
| ○ Neuroprotection | 3.2 | Animal Alzheimer and Parkinson models show effect; no confirmatory human data |
| ○ Cognition / Focus | 3.0 | Animal models suggest BBB penetration and mild neuroprotection; no human cognitive RCTs |
| ○ Cellular Senescence | 3.0 | Senolytic signals in preclinical only |
| ○ Energy / Fatigue | 3.0 | Metabolic improvement in dysregulated users; no energy data in healthy |
Frequently Asked Questions
How does berberine lower blood sugar?
Berberine activates AMPK via mild complex I inhibition, lowering HbA1c by 0.71% across 27 RCTs (Lan 2015 meta-analysis, n=2,569). The AMPK cascade suppresses hepatic gluconeogenesis (PEPCK, G6Pase downregulation), upregulates GLUT4 translocation, and inhibits PTP1B to restore IRS-1 signaling. A secondary pathway stimulates GLP-1 secretion from intestinal L-cells through microbiome modulation. Mechanistically analogous to metformin with overlapping downstream targets.
What is the correct berberine dose?
500 mg three times daily with meals (1.5 g/day total) is the RCT-validated dose used in Yin 2008 (PMID 18442638) and all major meta-analyses. Start 500 mg once daily for 1 week to assess GI tolerance, then titrate to TID. With meals is non-negotiable: fasting dose collapses glycemic effect and worsens GI symptoms. Dihydroberberine (Glucovantage) effective dose is 100-200 mg BID.
How strong is the berberine evidence versus metformin?
Yin 2008 (PMID 18442638, n=84) head-to-head showed berberine HbA1c −2.0% versus metformin −1.8% in newly diagnosed type 2 diabetes (non-inferior). Lan 2015 meta-analysis pooled 27 RCTs (n=2,569): HbA1c −0.71%, fasting glucose −1.21 mmol/L. Caveat: 90%+ of trials are Chinese single-center with small samples (<100) and variable blinding. No Cochrane review exists. Western replication is the key gap.
Is berberine safe long-term?
Short-term (3-6 months) safety is well characterized: GI events in 10-35% of users (LiverTox, 27 trials, n=2,569), no serious adverse events. Beyond 12 months, data is sparse. The EFSA 2026 draft opinion (NutraIngredients, March 2026) concluded no safe intake level can be established, citing in-vitro genotoxicity (gene mutation, chromosomal damage), idiosyncratic hepatotoxicity case reports, and rodent carcinogenicity. Public consultation closes May 4, 2026.
Who should not take berberine?
Hard contraindication in pregnancy and breastfeeding: berberine displaces bilirubin from albumin ~10x more potently than phenylbutazone (PMID 8513024), with documented neonatal kernicterus cases. Avoid with CYP3A4 substrates (most statins, macrolides, cyclosporine) after Guo 2012 (PMC4898966) showed midazolam AUC +40%, dextromethorphan metabolic ratio +9x, and cyclosporine AUC +34.5% in transplant patients. NIH MotherToBaby rates berberine unsafe perinatally.
How fast does berberine work?
Postprandial glucose changes are measurable within 3-7 days at 1.5 g/day (Yin 2008). Fasting glucose shifts at 2-4 weeks. HbA1c requires 8-12 weeks to reflect the 3-month erythrocyte turnover window. Lipid effects (LDL, triglycerides) plateau at 4-8 weeks per Koppen 2017 (PMID 28854908, 11 RCTs). NAFLD markers lag at 12-16 weeks. Gut antimicrobial effects (H. pylori, SIBO) emerge within 24-72 hours.
What is dihydroberberine (DHB) and is it better?
Dihydroberberine is a reduced form of berberine with ~5x higher plasma AUC at one-fifth the dose (Wang 2017, PMID 28836457). Effective dose is 100-200 mg BID versus 500 mg TID for standard HCl. Community reports fewer GI events. Caveat: all efficacy and safety RCTs (Yin 2008, Lan 2015, Koppen 2017) used HCl. DHB has pharmacokinetic validation but no long-term efficacy RCTs. Liposomal berberine is marketing-driven with thin PK data.
Is berberine really nature's Ozempic?
No. Berberine produces 1-2 kg weight loss in metabolically dysregulated populations (Peng 2023 meta, 12 RCTs, n=743): BMI −0.45 kg/m², waist −1.23 cm. Semaglutide produces 15-20% body weight loss (STEP trials). The effect-size difference is roughly 18-fold. In metabolically healthy users, berberine's weight effect approaches zero. AAFP, NCCIH, and OPSS published corrections in 2023-2024 rejecting the Ozempic equivalence.
How This Score Could Change
BioHarmony scores are living assessments. New research, regulatory changes, or personal context can shift the score up or down. These are the most likely scenarios that would change this intervention's rating.
| Scenario | Dimensions Affected | Recalculated Score | New Tier |
|---|---|---|---|
| EFSA finalizes Annex III Part A prohibition (May 2026+) | Safety: 2.5 → 4.0 raw; Evidence: 3.5 → 3.0 | ~5.3 | Neutral |
| Long-term (>24 mo) DHB RCT lands with maintained HbA1c and cleaner hepatic markers | Efficacy: 3.5 → 4.0 (DHB split); Evidence: 3.5 → 4.0; Safety: 2.5 → 2.0 on DHB only | ~7.6 (DHB variant) | Strong recommend |
| Western-led Phase 3 quality replication matches Lan 2015 effect sizes | Evidence: 3.5 → 4.0; Efficacy: 3.5 → 4.0 | ~7.5 | Strong recommend |
| Hepatotoxicity case cluster becomes a case series with identifiable risk factors | Safety: 2.5 → 3.5; Side Effects: 2.8 → 3.5 | ~5.6 | Neutral |
| GLP-1 upregulation via microbiome shift confirmed dose-responsively in humans | Breadth: 3.8 → 4.2; Evidence: 3.5 → 3.8 | ~7.2 | Worth trying (strengthened) |
| Major PBM or FDA upgrades berberine to "avoid" class for CYP interactions | Safety: 2.5 → 3.5; Opportunity: 1.5 → 2.5 | ~5.8 | Neutral |
BioHarmony v0.5 methodology. Date scored: 2026-04-17. Author: Nick Urban, CHEK Functional Health Coach Level 1.
Key Evidence Sources
- Lan J et al. 2015. Meta-analysis of 27 RCTs (n=2,569) in T2D and hyperlipidemia: HbA1c −0.71%, fasting glucose −1.21 mmol/L. Largest meta-analysis. Quality moderate; majority Chinese single-center trials
- Yin J, Xing H, Ye J. 2008. Head-to-head berberine vs metformin in newly diagnosed T2D (n=84): HbA1c −2.0% berberine, −1.8% metformin (non-inferior). Landmark head-to-head RCT
- Koppen LM et al. 2017. Lipid meta-analysis (11 RCTs, n=874): LDL −0.65 mmol/L, triglycerides −0.50 mmol/L, total cholesterol −0.61 mmol/L. PCSK9 mechanism explains LDL effect distinct from statins
- Wei W et al. 2012. PCOS meta-analysis: ovulation and androgen effects comparable to metformin. 3 RCTs, 535 women
- Yan HM et al. 2015. NAFLD RCT (n=184, 16 weeks): ALT −36%, steatosis grade improved in 63% berberine vs 31% control. Imaging-confirmed liver fat reduction
- Wang Y et al. 2017. Dihydroberberine pharmacokinetics: ~5x plasma AUC versus HCl. PK validated; efficacy RCTs absent
- Guo Y et al. 2012. In-human CYP inhibition at 300 mg TID for 2 weeks: CYP3A4 midazolam AUC +40%, CYP2D6 dextromethorphan ratio +9x, CYP2C9 losartan ratio doubled. Core drug interaction reference
- Wu X et al. 1995. Berberine displaces bilirubin from albumin ~10x more potently than phenylbutazone. Mechanism for pregnancy kernicterus contraindication
- EFSA NDA Panel. January 2026 draft scientific opinion: no safe intake level can be established for berberine in food supplements. Public consultation closes May 4, 2026. Genotoxicity, hepatotoxicity, carcinogenicity signals. EU Annex III Part A prohibition possible
- NIH MotherToBaby. Berberine in pregnancy fact sheet: rated unsafe in pregnancy and breastfeeding. Neonatal kernicterus risk
- Dong H et al. 2012. Berberine NAFLD systematic review and meta-analysis. 14 RCTs, n=1,068, HbA1c WMD −0.92%
- Chen YX et al. 2020. Lancet Gastroenterol Hepatol: berberine colorectal adenoma recurrence RCT. Longest-duration berberine RCT; 6-year follow-up
- NIH LiverTox. Berberine entry: GI adverse events in 10-35% across 27 trials (n=2,569); Likelihood Score E for hepatotoxicity. Safety summary aggregated from RCT literature
- Peng L et al. 2023. Frontiers Pharmacology weight meta-analysis (12 RCTs, n=743): BMI −0.45 kg/m2, waist −1.23 cm. Modest weight effect; dysregulated populations only
- NCCIH. Berberine and weight loss: official NIH position against Ozempic equivalence framing. Federal corrective on TikTok nature's Ozempic trend
Other interventions for Metabolic Health
See all ratings →📊 How BioHarmony scoring works
BioHarmony translates a weighted expected-value calculation into a reader-facing 0–10 score. 5.0 is neutral (benefits and risks balance). Above 5 = benefits outweigh risks; below 5 = risks outweigh benefits.
Harm-type downsides (safety risk, side effects, reversibility, dependency) carry a 1.4× precautionary multiplier. Harm weighs more than benefit. Opportunity-type downsides (financial cost, time/effort, opportunity cost) are subtracted at face value.
Use case subratings are independent assessments of how well the intervention addresses specific health goals. They are not components of the overall score. Each subrating reflects the scorer's judgment based on use-case-specific evidence, safety, and effect sizes.
Every dimension is evaluated on a 1–5 scale, and the baseline (1) is subtracted before weighting. A perfect intervention with zero downsides contributes zero penalty rather than a residual floor, so top-tier scores are actually reachable.
EV = Upside − Downside
EV = 2.270 − 1.160 = 1.110
EV ranges from −5 to +5. Adding 7 shifts to 2–12, dividing by 12 normalizes to 0–1, then ×10 gives the 0–10 score.
Score = ((1.110 + 7) / 12) × 10 = 6.8 / 10
Further learning
The ONLY 5 Top Berberine Supplements Review 2026: Nature’s Ozempic or Better?
Discover the incredible benefits of berberine, a potent plant extract used in ancient medical systems like Ayurveda and Chinese Medicine. Berberine has been shown…
Improve Gut Health to Protect Your Brain, Metabolism & Longevity
Martha Carlin, BiotiQuest founder, on why single-strain probiotics fail, horizontal gene transfer and antibiotic resistance, the glycocalyx as the real first line of gut…
