Curcumin
Evidence Fact Sheet
Curcuma longa · turmeric polyphenol
Curcumin is the principal polyphenol of turmeric (Curcuma longa), studied for NF-κB / NRF2 / COX-2 modulation of inflammatory and oxidative pathways. Research typically uses 500-2000 mg/day standardized 95% curcuminoids or lower-dose bioavailability-enhanced forms. Evidence spans inflammation, joints, metabolic and liver markers, with honest null findings too.
Also known as: Diferuloylmethane · Curcuma longa extract · Bioavailable curcumin (Meriva · Theracurmin · Longvida)
Overview
Curcumin is the principal curcuminoid in turmeric (Curcuma longa), a culinary spice polyphenol. Mechanistic research links it to NF-κB pathway inhibition, NRF2/ARE antioxidant-response activation, dual COX-2 and 5-LOX inhibition, and suppression of inflammatory cytokines such as TNF-α and IL-6, with most clinical interest centred on inflammatory, joint, metabolic, and hepatic markers. A central caveat is its low oral bioavailability, so research often uses lipid/phospholipid or piperine co-formulations (e.g. Meriva, Theracurmin, Longvida) at lower effective doses, while standardized extract studies typically use 500-2000 mg/day of 95% curcuminoids. Regulatory status: turmeric and curcumin are GRAS in the US and marketed under DSHEA as dietary supplements with structure/function claims only; EFSA set an ADI of 3 mg/kg bw/day (under re-evaluation) with no authorized health claim; ANVISA permits it as a dietary supplement in Brazil; and China treats turmeric as medicine-food homology and curcumin as food colour E100, with no authorized single-ingredient health function. This is an evidence summary, not medical or dosing guidance.
Mechanism of Action
NF-κB pathway inhibition · NRF2/ARE antioxidant response activation · COX-2 + 5-LOX dual inhibition · TNF-α + IL-6 suppression
Body systems: Immune System · Musculoskeletal · METABOLISM · Cardiovascular
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 — Curcumin is not characterized as a treatment for any disease.
Inflammation Markers (CRP, IL-6, TNF-α)
Meta-analysis supported- WMD -0.58 mg/lCRP · 95%CI -0.74,-0.41
- WMD -3.48 pg/mlTNF-α · 95%CI -4.38,-2.58
In a GRADE-assessed dose-response meta-analysis of 66 randomized controlled trials, turmeric/curcumin supplementation was associated with significantly lower inflammatory markers — CRP, TNF-α and IL-6 — and improved antioxidant status (higher TAC, lower MDA). The same analysis found NO significant change in IL-1β, an honest mixed signal within the inflammation panel.
Reported effect: CRP WMD -0.58 mg/l (95% CI -0.74, -0.41); TNF-α WMD -3.48 pg/ml (95% CI -4.38, -2.58); IL-6 WMD -1.31 pg/ml (95% CI -1.58, -0.67); IL-1β WMD -0.46 pg/ml (95% CI -1.18, 0.27) not significant
“Sixty-six RCTs were included in the final analysis. We observed that turmeric/curcumin supplementation significantly reduces levels of inflammatory markers, including CRP (WMD: -0.58 mg/l, 95 % CI: -0.74, -0.41), TNF-α (WMD: -3.48 pg/ml, 95 % CI: -4.38, -2.58), and IL-6 (WMD: -1.31 pg/ml, 95 % CI: -1.58, -0.67); except for IL-1β (WMD: -0.46 pg/ml, 95 % CI: -1.18, 0.27) for which no significant change was found.”
Source: PMID 36804260 · Dehzad 2023 · Cytokine
Joint Comfort / Osteoarthritis
Meta-analysis supported- MD -2.04pain VAS · 3 RCTs · p<.00001
- MD -15.36WOMAC · 4 RCTs · p=.009
A systematic review and meta-analysis of 8 RCTs found turmeric/curcumin (about 1000 mg/day curcumin) reduced pain VAS and WOMAC scores versus placebo. Notably, the abstract reported NO significant difference in pain VAS between turmeric/curcumin and pain medication, and the authors caution the trial count, sample size and methodological quality were not sufficient for definitive conclusions.
Reported effect: PVAS mean difference -2.04 [-2.85, -1.24] vs placebo (P<.00001); WOMAC mean difference -15.36 [-26.9, -3.77] (P=.009); no significant difference vs pain medicine
“Three among the included RCTs reported reduction of PVAS (mean difference: -2.04 [-2.85, -1.24]) with turmeric/curcumin in comparison with placebo (P < .00001), whereas meta-analysis of four studies showed a decrease of WOMAC with turmeric/curcumin treatment (mean difference: -15.36 [-26.9, -3.77]; P = .009). Furthermore, there was no significant mean difference in PVAS between turmeric/curcumin and pain medicine in meta-analysis of five studies.”
Source: PMID 27533649 · Daily 2016 · J Med Food
Glycemic Control (Type 2 Diabetes)
Meta-analysis supported- 18 trials · 1382T2D participants
- MD -11.48 mg/dLfasting glucose · p<0.01
- MD -0.54%HbA1c · p<0.01
A meta-analysis of 18 trials in 1382 people with type 2 diabetes found curcumin supplementation significantly lowered fasting blood glucose and glycated hemoglobin, plus C-reactive protein, versus placebo. The authors note inconsistency across individual trials and recommend better dosing/absorption formulations in future work.
Reported effect: Fasting blood glucose MD -11.48 mg/dL (95% CI -14.26, -8.70), p<0.01; HbA1c MD -0.54% (95% CI -0.73, -0.35), p<0.01; CRP SMD -0.59 (95% CI -1.11, -0.07), p=0.03
“The evidence from 18 trials with 1382 T2D with a mean age of 55.9 years was analyzed. Supplementation with curcumin led to a significant decrease in fasting blood glucose, MD = -11.48 mg/dL, 95%CI (-14.26, -8.70), p < 0.01 and glycated hemoglobin, MD = -0.54%, 95%CI (-0.73, -0.35), p < 0.01.”
Source: PMID 39683570 · Mokgalaboni 2024 · Nutrients
Liver Markers (NAFLD)
Meta-analysis supported- 11 MAs · 5546umbrella · participants
- ES -1.072AST · p=0.000
- ES -0.625ALT · p=0.014
An umbrella meta-analysis pooling 11 meta-analyses of 99 RCTs (5546 participants) found curcumin/turmeric significantly reduced the liver enzymes AST and ALT, plus HOMA-IR, BMI and waist circumference, in NAFLD patients. Honest negatives within the same analysis: effects on GGT, ALP, total cholesterol, LDL-C, HDL-C, fasting blood sugar, HbA1c and body weight were NOT significant.
Reported effect: AST ES -1.072 (95% CI -1.656, -0.488), p=0.000; ALT ES -0.625 (95% CI -1.170, -0.134), p=0.014; HOMA-IR ES -0.291, p=0.000; effects on GGT, ALP, TC, LDL-C, HDL-C, FBS, HbA1C and body weight not significant
“Findings of 11 meta-analyses of 99 randomized controlled trials comprising 5546 participants revealed that curcumin/turmeric supplementation reduced AST (ES = -1.072, 95% CI (-1.656, -0.488), p = 0.000), ALT (ES = -0.625, 95% CI (-1.170, -0.134), p = 0.014) ... However, the effects of curcumin on GGT, ALP, TC, LDL-C, HDL-C, FBS, and HbA1C levels and body weight were not significant.”
Source: PMID 37918958 · Molani-Gol 2024 · Phytother Res
Depression and Anxiety Symptoms
Meta-analysis supported- Hedge's g -0.75depression · 10 studies · 531
- Hedge's g -2.62anxiety · 5 studies · 284
- p<0.001both pooled effects
A PRISMA meta-analysis of 9 eligible trials reported a significant overall effect of curcumin on depressive symptoms and on anxiety symptoms, with a large pooled effect size. The authors explicitly caution that the small sample size means results should be interpreted carefully, and most trials added curcumin to standard care rather than testing it alone.
Reported effect: Depression: 10 studies, 531 participants, Hedge's g = -0.75 (95% CI -1.11 to -0.39, p<0.001); Anxiety: 5 studies, 284 participants, Hedge's g = -2.62 (95% CI -4.06 to -1.17, p<0.001)
“We found an overall significant effect of curcumin on depressive (10 studies, 531 participants, Hedge's g = -0.75, 95% CI -1.11 to -0.39, p < 0.001) and anxiety symptoms (5 studies, 284 participants, Hedge's g = -2.62, 95% CI -4.06 to -1.17, p < 0.001), with large effect size.”
Source: PMID 31423805 · Fusar-Poli 2020 · Crit Rev Food Sci Nutr
Blood Pressure and Cardiovascular Risk Markers
RCT supported- 72 patientsT2DM · ASCVD ≥5%
- P<=0.001 / P=0.020SBP / DBP reduced
- P=0.004ASCVD risk class improved
In a randomized controlled trial of 72 diabetic patients with elevated atherosclerotic risk, curcumin added to conventional therapy significantly lowered systolic and diastolic blood pressure, LDL-C, TNF-α and MDA, raised HDL-C, and improved ASCVD risk classification. Honest negatives: there were NO significant between-group differences in HbA1c, fasting blood glucose, total cholesterol, triglycerides, BMI or heart rate.
Reported effect: n=72; SBP reduced P≤0.001, DBP P=0.020; ASCVD risk classification improved P=0.004; LDL-C P=0.024, TNF-α P=0.044, MDA P=0.028 decreased, HDL-C increased P=0.024; no significant difference in HbA1c, FBG, TC or TG (P>0.05)
“Seventy-two diabetic patients with an ASCVD risk score of ≥ 5% were randomly assigned to Curcumin group ... Curcumin significantly reduced SBP and DBP (P ≤ 0.001 and P = 0.020, respectively) and improved ASCVD risk classification (P = 0.004). LDL-C (P = 0.024), TNF-α (P = 0.044), and MDA (P = 0.028) levels decreased, while HDL-C increased (P = 0.024) versus control. No significant differences were found between groups regarding HbA1c, FBG, TC or TG (P > 0.05).”
Source: PMID 40759997 · El-Rakabawy 2025 · Sci Rep
Brain-Derived Neurotrophic Factor (BDNF)
Meta-analysis supported- WMD 1789.38 pg/mLserum BDNF · P<.01
- 4 RCTs · 139trials · participants
- I2 = 83.5%high heterogeneity
A dose-response meta-analysis of 4 RCTs (139 participants) found short-term curcumin supplementation significantly increased serum BDNF, a neurotrophin linked to learning and memory. The signal carries substantial heterogeneity (I2 = 83.5%) and rests on a small evidence base, so it is a mechanistic biomarker finding rather than a demonstrated cognitive benefit.
Reported effect: Serum BDNF weighted mean difference 1789.38 pg/mL (95% CI 722.04-2856.71, P<.01), 4 RCTs, 139 participants, I2=83.5%
“Four randomized control trials with 139 participants were included. ... Curcumin supplementation significantly increased serum BDNF levels (weighted mean difference: 1789.38 pg/mL, 95% confidence interval: 722.04-2856.71, P < .01) with significant heterogeneity among the studies (I2 = 83.5%, P < .001).”
Source: PMID 31279955 · Sarraf 2019 · Nutr Res
Cognitive Decline in Alzheimer's Models (Honest Negative — Animal Only)
Null / no benefit Emerging / indexed- 33 articlesanimal AD models
- -2.027acquisition effect · p<0.001
- publication biasdetected
Despite popular framing, the strongest pooled Alzheimer's cognition evidence is in ANIMAL models, not humans: a meta-analysis of animal AD studies found curcumin improved maze acquisition and retention, but the authors explicitly note the cognitive benefit was recently controversial, that publication bias was detected, and that well-designed studies are still needed before clinical conclusions. This is an honest preclinical-only signal.
Reported effect: Animal AD models: acquisition global effect -2.027 (95% CI -2.435 to -1.619, p<0.001); retention 1.606 (95% CI 1.101 to 2.111, p<0.001); 33 articles, 29 in meta-analysis; publication bias detected
“Our systematic review included 33 articles. A meta-analysis of 29 publications showed that curcumin exerts significant positive effects on cognitive performance. For acquisition, the global estimated effect of curcumin was - 2.027 (95% CI - 2.435 to - 1.619, p < 0.001) ... Additionally, publication bias was detected. We conclude that curcumin may reduce cognitive deficits in experimental AD.”
Source: PMID 38265680 · Fan 2024 · Naunyn Schmiedebergs Arch Pharmacol
Oral Bioavailability / Formulation
RCT supported- 45.9-foldCHC vs unformulated
- 7.9-foldphytosome vs unformulated
- p<0.001absorption difference
Standard curcumin is poorly absorbed, which is the central practical limitation behind enhanced formulations. In a randomized double-blind crossover human study, total curcuminoid appearance in blood was higher for a phytosome formulation and far higher for a hydrophilic-carrier formulation versus unformulated standard curcumin, illustrating why research doses are not interchangeable across products.
Reported effect: Total curcuminoids in blood: phytosome (CP) 7.9-fold higher and turmeric volatile-oil (CTR) 1.3-fold higher vs unformulated standard; hydrophilic-carrier (CHC) 45.9-fold higher than standard (p<0.001)
“Total curcuminoids appearance in the blood was 1.3-fold higher for CTR and 7.9-fold higher for CP in comparison to unformulated CS. CHC showed a 45.9-fold higher absorption over CS and significantly improved absorption over CP (5.8-fold) and CTR (34.9-fold, all p < 0.001).”
Source: PMID 24461029 · Jäger 2014 · Nutr J
Overall Human-Health Scope and Certainty (Honest Negative — Mixed)
Null / no benefit Meta-analysis supported- 103 RCTs · 7216trials · participants
- 23 of 42outcomes significant (55%)
The most comprehensive synthesis pooled 103 RCTs across 42 outcomes in 7216 participants and found that only about 55% of outcomes reached statistical significance, with GRADE high-certainty evidence for just four (fasting blood sugar, CRP, HDL, weight). The authors stress that methodological quality was limited and that well-designed, long-term, large RCTs are still needed — a balanced reality check against over-claiming.
Reported effect: 103 RCTs on 42 outcomes, 7216 participants; 23/42 (55%) outcomes statistically significant; GRADE high certainty for FBS, CRP, HDL and weight only
“In total, 103 RCTs on 42 outcomes were included, incorporating a total population of 7216 participants. Overall, 23 out of 42 (55%) outcomes reported statistically significant effect sizes. The credibility of the evidence was rated as high for fasting blood sugar (FBS), C-reactive protein (CRP), high-density lipoprotein (HDL), and weight.”
Source: PMID 39478418 · Jafari 2024 · Phytother Res
Dosage (research context · not a recommendation)
500-2000 mg/day standardized curcumin (95% curcuminoids); bioavailability-enhanced (Meriva 1000 mg/day · Theracurmin 90-180 mg/day) lower effective dose
Regulatory Status · 4 Markets
- US · FDA
- GRAS (turmeric · curcumin) · DSHEA dietary supplement; structure/function claims
- EU · EFSA
- ADI 3 mg/kg bw/day (2010) under re-evaluation 2024; no authorized health claim
- CN · China
- China SAMR: turmeric is medicine-food homology; curcumin is approved food-colour additive E100 (GB 2760) and a permitted health-food raw material, but no single-ingredient curcumin authorized health function exists.
- BR · ANVISA
- RDC 243/2018 dietary supplement · IN 28/2018 alegação funcional not authorized
Safety
Low oral bioavailability (lipid/phospholipid/piperine co-formulation needed); rare hepatotoxicity case reports at high-dose extract (FDA MedWatch · 2023 EFSA opinion); avoid in gallstone obstruction
Related
Goals: inflammation-relief · joint-bone · heart-health · longevity-stack
Lifestyles: senior-60-plus · high-stress
References
PubMed-indexed citations anchoring the benefit findings above. Effect sizes are reported as published.
- PMID 36804260 · Dehzad 2023 · Cytokine — Inflammation Markers (CRP, IL-6, TNF-α)
- PMID 27533649 · Daily 2016 · J Med Food — Joint Comfort / Osteoarthritis
- PMID 39683570 · Mokgalaboni 2024 · Nutrients — Glycemic Control (Type 2 Diabetes)
- PMID 37918958 · Molani-Gol 2024 · Phytother Res — Liver Markers (NAFLD)
- PMID 31423805 · Fusar-Poli 2020 · Crit Rev Food Sci Nutr — Depression and Anxiety Symptoms
- PMID 40759997 · El-Rakabawy 2025 · Sci Rep — Blood Pressure and Cardiovascular Risk Markers
- PMID 31279955 · Sarraf 2019 · Nutr Res — Brain-Derived Neurotrophic Factor (BDNF)
- PMID 38265680 · Fan 2024 · Naunyn Schmiedebergs Arch Pharmacol — Cognitive Decline in Alzheimer's Models (Honest Negative — Animal Only)
- PMID 24461029 · Jäger 2014 · Nutr J — Oral Bioavailability / Formulation
- PMID 39478418 · Jafari 2024 · Phytother Res — Overall Human-Health Scope and Certainty (Honest Negative — Mixed)
Frequently Asked Questions
1. What does the strongest evidence actually show for curcumin?
The best-supported findings are on inflammatory and metabolic biomarkers. A meta-analysis of 66 RCTs reported lower CRP, TNF-α and IL-6, and the largest synthesis (103 RCTs, 7216 participants) rated certainty as high specifically for fasting blood sugar, CRP, HDL and weight. These are research findings on markers in studied populations, not disease treatment claims.
2. Why is curcumin's low bioavailability such a big deal?
Unformulated curcumin is poorly absorbed. In a crossover human study, blood curcuminoid levels were 7.9-fold higher for a phytosome formulation and 45.9-fold higher for a hydrophilic-carrier formulation versus standard curcumin. That is why research doses differ widely between plain 95% curcuminoid extracts and enhanced products like phospholipid (Meriva), Theracurmin or Longvida forms, and why product results are not interchangeable.
3. Does curcumin help with joint discomfort?
A meta-analysis of 8 RCTs (around 1000 mg/day curcumin) found reductions in pain VAS (MD -2.04) and WOMAC scores (MD -15.36) versus placebo. Importantly, the same analysis found no significant difference in pain VAS between curcumin and pain medication, and the authors caution the trials were too few and small for definitive conclusions. This is osteoarthritis research context, not a treatment recommendation.
4. What are the honest negatives for curcumin?
Several. The 66-RCT inflammation meta-analysis found no significant change in IL-1β. The NAFLD umbrella analysis found no significant effect on GGT, ALP, total cholesterol, LDL-C, HDL-C, fasting blood sugar, HbA1c or body weight. The cardiovascular RCT showed no significant change in HbA1c, fasting glucose, total cholesterol or triglycerides. And the strongest Alzheimer's cognition meta-analysis is in animal models with publication bias detected — not human disease evidence.
5. Is the Alzheimer's / cognition story real?
The cleanest human signal is a biomarker one: a 4-RCT meta-analysis found higher serum BDNF, but with high heterogeneity (I2 = 83.5%) and a small evidence base. The pooled Alzheimer's cognition benefit comes from animal models (33 studies) where publication bias was detected. So curcumin and cognition is best described as emerging and mechanistic, not an established human benefit.
6. Is curcumin approved as a treatment anywhere?
No. Turmeric and curcumin are GRAS in the US and sold under DSHEA with structure/function claims only; EFSA set an ADI of 3 mg/kg bw/day (under re-evaluation) with no authorized health claim; Brazil's ANVISA permits it as a dietary supplement; and China treats turmeric as medicine-food homology and curcumin as food colour E100, with no authorized single-ingredient health function. Everything above is evidence reporting, not medical, dosing or treatment advice.
Last evidence review: 2026-05-29