Ginkgo

What it is

Ginkgo is a herbal preparation made from the leaves (and, in food use, the seeds) of the tree Ginkgo biloba. Supplement products are usually standardised dried leaf extracts such as EGb 761, which are characterised by their flavonol glycoside and terpene lactone content. The leaf extract is marketed for brain and circulatory health, and neuroprotective effects are considered biologically plausible. Analytical surveys show that commercial ginkgo leaf extracts are a frequent target of economically motivated adulteration with cheaper flavonols such as rutin and quercetin or with extracts of Styphnolobium japonicum and buckwheat.

What the research says

The strongest current synthesis is a 2026 Cochrane review of 82 randomised trials in 10,613 people. It found that in people with mild cognitive impairment ginkgo probably has little or no effect at six months, while in people already diagnosed with dementia there may be small to moderate benefits on global status, cognition and daily activities, at low certainty and with very high statistical heterogeneity. The two large US Ginkgo Evaluation of Memory (GEM) trials, which followed 3,069 older adults for a median 6.1 years, found that ginkgo neither reduced the incidence of dementia nor slowed cognitive decline. Harms reported in trials are mostly mild, but a two-year National Toxicology Program rodent study found liver and thyroid tumours, and ginkgo seeds are toxic.

Evidence grade: Disputed.

What goes wrong

In people with subjective cognitive complaints, the Cochrane review found ginkgo may raise the risk of adverse events at three months, with uncertain benefit. (Source 1)

• Systematic review, Very low certainty.
• Size: 3 studies, 597 participants in this comparison.
• Who: People with subjective cognitive complaints but no diagnosis.
• How long: Three to six months.
• Result: Global clinical status MD 0.00, 95% CI -0.33 to 0.33, P = 1.00, 1 study, 197 participants, very low-certainty evidence; a larger three-month study found a possibly higher risk of adverse events.
• Funding: not stated.

Limit of this finding: The figures quoted for this group are internally consistent, but the same review contains a printed result that is not: for people who already have dementia it gives "MD -0.06, 95% CI -1.00 to -0.20", a point estimate outside its own confidence interval. That error does not touch this finding. Separately, the review is explicit that it does not know whether ginkgo helps this group at all, and the signal of more adverse events rests on a single three-month study, so this should be read as an unresolved question rather than a demonstrated harm.

In people with cognitive complaints, we are unsure whether ginkgo improves global clinical status at six months, and it may be associated with an increased risk of AEs at three months.

A two-year National Toxicology Program gavage study found ginkgo biloba extract caused liver and thyroid tumours in rodents. (Source 2)

• Animal study, Certainty not rated.
• Size: B6C3F1/N mice and F344/N rats, 3-month and 2-year studies.
• Who: Rodents (not humans)
• How long: 3 months and 2 years.
• Result: High incidence of hepatoblastomas in male and female mice; evidence of carcinogenic potential in the thyroid gland of both mice and rats; non-neoplastic lesions in liver, thyroid and nose.
• Funding: independent (US National Toxicology Program)

Key findings included a notably high incidence of hepatoblastomas in male and female mice and evidence of carcinogenic potential in the thyroid gland of both mice and rats.

An industry-funded adulteration bulletin reports that four of 14 commercial ginkgo products bought in the Edmonton, Alberta area were likely adulterated with pure flavonols rather than genuine ginkgo leaf extract - a separate survey from the Beltsville, Maryland sample described elsewhere in the same bulletin. (Source 3)

• Expert review, not systematic, Low certainty.
• Size: 14 commercial ginkgo products in the Edmonton survey.
• Who: Commercial ginkgo leaf products sold to the public, not people.
• How long: Not applicable; a one-off analysis of purchased products.
• Result: Four of 14 products likely adulterated with pure flavonols (rutin, quercetin, kaempferol and isorhamnetin); the bulletin does not give a certainty rating.
• Funding: American Botanical Council Botanical Adulterants Prevention Program, supported by industry and trade sponsors.

four out of 14 commercial ginkgo products sourced in the Edmonton, Alberta (Canada) area were likely adulterated with pure flavonols (rutin, quercetin, kaempferol, and isorhamnetin)

The same bulletin reports a separate analysis in which three of eight products labelled as containing ginkgo extract had chromatographic profiles resembling extracts of Japanese sophora, a different plant. (Source 4)

• Expert review, not systematic, Low certainty.
• Size: 8 products labelled to contain ginkgo extracts.
• Who: Commercial products labelled as ginkgo extract, not people.
• How long: Not applicable; a one-off analysis of purchased products.
• Result: Three of eight ginkgo-labelled products matched commercial Japanese sophora extracts on chromatographic profile; no clinical outcome was measured.
• Funding: American Botanical Council Botanical Adulterants Prevention Program, supported by industry and trade sponsors.

chromatographic profiles of three out of eight products labeled to contain ginkgo extracts closely resembled those of commercial extracts obtained from Japanese sophora

The same bulletin separately describes a 2012 analysis by Harnly and colleagues of 18 ginkgo supplements bought in the Beltsville, Maryland area or online, of which seven were judged clearly adulterated - a different sample from the 14-product Edmonton survey. (Source 5)

• Expert review, not systematic, Low certainty.
• Size: 18 ginkgo dietary supplements in the Harnly et al. analysis, of which seven were judged clearly adulterated.
• Who: Commercial ginkgo supplements purchased in the Beltsville, MD area or from the internet, not people.
• How long: Not applicable; a one-off analysis of purchased products.
• Result: Seven of 18 products clearly adulterated with either rutin, quercetin or an unidentified flavonol glycoside, judged by comparison of chromatographic and spectral fingerprints with authentic ginkgo extract; this is a separate sample from the Edmonton finding, so the counts must not be added together.
• Funding: American Botanical Council Botanical Adulterants Prevention Program, supported by industry and trade sponsors.

Also in 2012, Harnly et al. analyzed 18 ginkgo dietary supplements purchased in the Beltsville, MD area or from the internet. Comparison of chromatographic and spectral fingerprints with authentic ginkgo extract led to the conclusion that seven products were clearly adulterated with either rutin, quercetin, or an unidentified flavonol glycoside.

A second analytical programme states that adulteration of Ginkgo biloba plant material and extracts is becoming an increasingly common practice. (Source 6)

• Lab study in cells, Certainty not rated.
• Size: 32 samples of Ginkgo biloba plant materials and dried leaf extracts.
• Who: Plant materials and dried leaf extracts.
• How long: not applicable.
• Result: Non-targeted LC-HRMS with chemometrics used to distinguish authentic from non-authentic material; no single adulteration rate given for the 32 samples.
• Funding: independent (US National Institute of Standards and Technology)

Adulteration of Ginkgo biloba L. plants and extracts are becoming more and more common practice due to economically driven motivation from increasing demand in the market and the high cost of raw materials and production.

The US National Center for Complementary and Integrative Health's position, current at February 2025, is that there is no conclusive evidence of benefit for any condition and that ginkgo seeds are toxic. (Source 7)

• Official position, Certainty not rated.
• Size: not applicable.
• Who: General public.
• How long: Page last updated February 2025.
• Result: Agency states no conclusive benefit; names dizziness, gastrointestinal symptoms and headache as the commonest side effects; warns of bleeding risk with anticoagulants such as warfarin and of possible harm in pregnancy.
• Funding: US government agency.

Fresh ginkgo seeds are toxic when consumed orally, and serious side effects have also occurred in people who consumed roasted ginkgo seeds or the crude ginkgo plant.

What the evidence supports

In people already diagnosed with dementia, a 2026 Cochrane review reported possible small to moderate six-month benefits on global status, cognition and daily activities, at low certainty and with very high heterogeneity; the size of the global-status benefit cannot be read from the review's published figures, which are internally inconsistent. (Source 1)

• Systematic review, Low certainty.
• Size: 13 studies, 3,288 participants in the dementia comparison (82 studies, 10,613 participants overall)
• Who: People with a diagnosis of dementia.
• How long: Six months for the main outcomes.
• Result: Global clinical status on a six-point Likert scale, printed in the review as MD -0.06, 95% CI -1.00 to -0.20 (these figures are internally inconsistent - see caveat), I2 = 88%, 5 studies, 1359 participants; Syndrom-Kurztest cognition MD -1.86, 95% CI -3.48 to -0.24, I2 = 96%, 9 studies, 2801 participants; ADL International Scale MD -0.19, 95% CI -0.35 to -0.03, I2 = 91%, 8 studies, 2571 participants.
• Funding: not stated.

Limit of this finding: The review's own abstract prints the dementia global-status result as "MD -0.06, 95% CI -1.00 to -0.20". That cannot be right: a point estimate has to sit inside its own confidence interval, and -0.06 does not sit inside -1.00 to -0.20. At least one of those three numbers is a misprint. We have quoted the review exactly as it is printed rather than guessing at what was meant. What this means for a reader: you can take from this result that the dementia trials pointed towards a benefit on global status, but you cannot take any number from it - not -0.06, and not anything worked out from the interval -1.00 to -0.20 - as the measured size of that benefit. The cognition and daily-activity results quoted alongside it are internally consistent and are not affected. The heterogeneity between trials was very high (I2 = 88% to 96%) and the review rates this evidence low certainty, so the benefit is uncertain in any case. Anyone wanting a figure for global status should go to the full review text or wait for a correction.

In people with dementia, there may be small to moderate benefits at six months for global status, cognition, and ADLs.

What the evidence does not support

In people with mild cognitive impairment, the same Cochrane review found ginkgo probably has little or no effect at six months on global status, cognition or activities of daily living. (Source 1)

• Systematic review, Moderate certainty.
• Size: 12 studies, 1,913 participants.
• Who: People with mild cognitive impairment.
• How long: Six months.
• Result: Clinical Dementia Rating MD -0.03, 95% CI -0.06 to 0.01, 3 studies, 631 participants, I2 = 0%; ADAS-cog MD -0.07, 95% CI -0.67 to 0.51, 2 studies, 508 participants; Instrumental ADL MD -0.05, 95% CI -0.29 to 0.19, 1 study, 350 participants.
• Funding: not stated.

Limit of this finding: Nothing in this finding depends on the faulty number, but a reader should know it is there: elsewhere in the same review, the result for people who already have dementia is printed as "MD -0.06, 95% CI -1.00 to -0.20", where the point estimate lies outside its own confidence interval and so cannot be correct. The mild-cognitive-impairment figures quoted here are internally consistent and stand on their own. It does mean the review's printed numbers are worth checking against the full text before any of them are relied on.

In people with MCI, ginkgo probably has little or no effect at six months on global status, cognition, or ADLS.

In the large GEM randomised trial, ginkgo 120 mg twice daily did not reduce the incidence of dementia or Alzheimer disease over a median 6.1 years. (Source 8)

• Randomized trial, High certainty.
• Size: 3,069 community volunteers (1,545 ginkgo, 1,524 placebo)
• Who: Community volunteers aged 75 years or older with normal cognition (n = 2587) or mild cognitive impairment (n = 482)
• How long: Median follow-up 6.1 years.
• Result: All-cause dementia HR 1.12, 95% CI 0.94-1.33, P = .21; Alzheimer disease HR 1.16, 95% CI 0.97-1.39, P = .11; progression to dementia in MCI HR 1.13, 95% CI 0.85-1.50, P = .39.
• Funding: not stated in the abstract; the extract was supplied for a US National Institutes of Health funded trial.

In this study, G. biloba at 120 mg twice a day was not effective in reducing either the overall incidence rate of dementia or AD incidence in elderly individuals with normal cognition or those with MCI.

The GEM trial's cognitive outcomes paper found no slowing of cognitive decline in any tested domain. (Source 9)

• Randomized trial, High certainty.
• Size: 3,069 participants aged 72 to 96 years.
• Who: Community-dwelling older adults with normal cognition or mild cognitive impairment.
• How long: Median follow-up 6.1 years.
• Result: Annual z-score decline did not differ, e.g. memory 0.043 (95% CI 0.034-0.051) vs 0.041 (95% CI 0.032-0.050); 3MSE P = .71; ADAS-Cog P = .97.
• Funding: not stated in the abstract.

Compared with placebo, the use of G. biloba, 120 mg twice daily, did not result in less cognitive decline in older adults with normal cognition or with mild cognitive impairment.

A meta-analysis of 18 randomised trials of standardised ginkgo extract found no measurable increase in bleeding risk on haemostasis outcomes. (Source 10)

• Meta-analysis, Low certainty.
• Size: 1,985 adults across 18 randomised controlled trials.
• Who: 87% patients with dementia, peripheral artery disease or diabetes; 13% healthy volunteers.
• How long: varied by trial.
• Result: ADP-induced platelet aggregation WMD -0.35%, 95% CI -15.16-14.46%; fibrinogen GIV -2.45 mg/dl, 95% CI -11.59-6.70; aPTT GIV -0.42 sec, 95% CI -0.97-0.12; PT SMD 0.00, 95% CI -0.09-0.09; blood viscosity reduced WMD -1.03 mPa.sec, 95% CI -1.29 to -0.78.
• Funding: not stated.

Based on meta-analysis of hemostasis outcomes, comparison of mean difference or baseline change between treatment and placebo groups did not indicate a higher bleeding risk associated with standardized GBE.

Where the research disagrees

Whether ginkgo helps people who already have dementia

• Cochrane review authors (Wieland and colleagues, 2026), systematic-review of 13 placebo-controlled trials, low-certainty evidence with I2 up to 96%: In people with dementia, there may be small to moderate benefits at six months for global status, cognition, and ADLs. (Source 1)
• US National Center for Complementary and Integrative Health (February 2025), position statement: Ginkgo has not been shown to be beneficial for preventing or slowing the progression of dementia. (Source 7)

Whether ginkgo increases bleeding risk

• Kellermann and Kloft (2011) meta-analysis, meta-analysis of 18 randomised trials, 1,985 adults: Based on meta-analysis of hemostasis outcomes, comparison of mean difference or baseline change between treatment and placebo groups did not indicate a higher bleeding risk associated with standardized GBE. (Source 10)
• US National Center for Complementary and Integrative Health (February 2025), position statement based on case reports and pharmacology: Ginkgo may increase the risk of bleeding in people who are taking anticoagulant drugs, such as warfarin. (Source 7)

How much

• Reference intake: No reference intake (RDA or AI) exists for ginkgo because it is not an essential nutrient; it is a herbal preparation, and the US NCCIH describes it as a supplement whose benefit is not established (page last updated February 2025). (Source 7)
• Upper limit: No tolerable upper intake level or acceptable daily intake has been set for ginkgo leaf extract by the bodies we could reach; NCCIH states only that moderate oral amounts of leaf extract are likely safe for most adults, while fresh and roasted seeds are toxic (February 2025). (Source 7)
• Studied: The GEM trial gave 120 mg of G. biloba extract twice daily (240 mg/day) for a median of 6.1 years. (Source 8)
• Studied: The Cochrane review pooled trials of ginkgo against placebo at six months in dementia, mild cognitive impairment, multiple sclerosis and subjective cognitive complaints, across 82 studies and 10,613 participants. (Source 1)

A common belief, and what the research shows

The belief: Taking ginkgo protects an ageing brain from dementia.

What the research shows: The largest randomised test of exactly that idea found the opposite of protection. Over a median 6.1 years in 3,069 people aged 75 or older, the GEM trial reported that G. biloba at 120 mg twice a day was not effective in reducing either the overall incidence rate of dementia or AD incidence in elderly individuals with normal cognition or those with MCI. Its companion cognitive paper reported that the use of G. biloba, 120 mg twice daily, did not result in less cognitive decline in older adults with normal cognition or with mild cognitive impairment. The 2026 Cochrane review's possible benefit applies to people who already have dementia, not to prevention, and is rated low certainty.

Questions and answers

What is it?

Ginkgo is a herbal product made from the leaves and seeds of the Ginkgo biloba tree. Supplements almost always contain a standardised dried leaf extract, not the raw leaf. The extract is characterised by flavonol glycosides and terpene lactones. (Source 11)

What does it do in the body?

Ginkgo is taken with the intention of improving brain and circulatory health, and laboratory work makes a neuroprotective effect plausible. What it actually does in people is much less certain: the largest randomised trials found no effect on dementia incidence or cognitive decline, while a 2026 Cochrane review found possible small benefits only in people who already have dementia. (Source 1)

Is it good or bad for you?

It depends on the form and the person. Leaf extract in moderate oral amounts is generally well tolerated in trials, with dizziness, gastrointestinal symptoms and headache the commonest complaints, but the US NCCIH position is that no benefit is conclusively established. Ginkgo seeds and crude plant material are a different matter and have caused serious poisoning. A two-year rodent study also found liver and thyroid tumours at high doses. (Source 7)

How do you get more of it?

Ginkgo is not present in an ordinary diet, so intake comes from leaf extract supplements or from ginkgo seeds eaten as food in parts of East Asia. The dose used in the largest randomised trial was 120 mg of extract twice a day. This is a description of what trials gave people, not a recommendation. (Source 8)

If it is harmful, what reduces it?

Nothing needs to be cleared from the body under normal use; ginkgo is not stored as a nutrient and exposure ends when supplementation stops. The relevant harms are dose-dependent and situational, so stopping the product, and avoiding ginkgo seeds and crude plant material, is what removes the exposure. (Source 7)

Why might someone be low in it or missing it?

This question does not apply. Ginkgo is not an essential nutrient or a body constituent, so nobody can be deficient in it; people simply either take a ginkgo product or they do not. The literature we searched treats it purely as an optional herbal preparation. (Source 1)

Which whole foods contain it or feed it?

The only food source is the ginkgo seed (often sold as ginkgo nut), eaten in parts of East Asia. This is not an equivalent of the leaf extract used in trials, and it carries its own toxicity: fresh seeds are toxic and serious effects have also followed eating roasted seeds or crude plant material. The leaf extract itself is not a food. (Source 7)

What happens if you do not have it?

Nothing is known to go wrong if you never take ginkgo, because it plays no role in normal physiology. In the largest placebo-controlled test, people who took placebo for a median of 6.1 years did not develop dementia any faster than those taking ginkgo, and their cognitive decline was no steeper. (Source 8)

How can you test for it?

There is no clinical test for ginkgo status, because there is no normal body level of it to measure. What can be tested is the product: analytical chemistry methods such as LC-HRMS with chemometrics and flavonol glycoside ratio checks are used to tell authentic ginkgo extract from adulterated material, and these surveys repeatedly find adulterated products on the market. (Source 6)

We searched: Searched for ginkgo biomarkers, ginkgo status testing and ginkgo authentication assays; found only product-authentication chemistry (NIST LC-HRMS study, American Botanical Council adulterants bulletin), no validated human test.

References

1. Cochrane Database of Systematic Reviews (Wieland LS, Ludeman E, Chi Y, Feinberg TM, Chen IH, Chen KH, Zhu Y, Wolverson E, Amri H). Ginkgo biloba for cognitive impairment and dementia. 2026. PMID 41641880, DOI 10.1002/14651858.CD013661.pub2. Read the source
2. Toxicologic Pathology (National Toxicology Program). Toxicity and Carcinogenicity Studies of Ginkgo biloba Extract in Rat and Mouse: Liver, Thyroid, and Nose Are Targets. 2014. PMID 23960164, DOI 10.1177/0192623313501235. Read the source
3. American Botanical Council, Botanical Adulterants Prevention Program (Stefan Gafner). Adulteration of Ginkgo biloba Leaf Extract - Botanical Adulterants Bulletin. 2018. Read the source
4. American Botanical Council, Botanical Adulterants Prevention Program (Stefan Gafner). Adulteration of Ginkgo biloba Leaf Extract - Botanical Adulterants Bulletin. 2018. Read the source
5. American Botanical Council, Botanical Adulterants Prevention Program (Stefan Gafner). Adulteration of Ginkgo biloba Leaf Extract - Botanical Adulterants Bulletin. 2018. Read the source
6. US National Institute of Standards and Technology (author manuscript). A nontargeted approach to determine the authenticity of Ginkgo biloba L. plant materials and dried leaf extracts by liquid chromatography-high resolution mass spectrometry (LC-HRMS) and chemometrics. 2023. Read the source
7. US National Center for Complementary and Integrative Health (NCCIH), National Institutes of Health. Ginkgo: Usefulness and Safety. 2025. Read the source
8. JAMA. Ginkgo biloba for prevention of dementia: a randomized controlled trial. 2008. PMID 19017911, DOI 10.1001/jama.2008.683. Read the source
9. JAMA. Ginkgo biloba for preventing cognitive decline in older adults: a randomized trial. 2009. PMID 20040554, DOI 10.1001/jama.2009.1913. Read the source
10. Pharmacotherapy. Is there a risk of bleeding associated with standardized Ginkgo biloba extract therapy? A systematic review and meta-analysis. 2011. PMID 21923430, DOI 10.1592/phco.31.5.490. Read the source
11. National Institute of Diabetes and Digestive and Kidney Diseases, NCBI Bookshelf. Ginkgo - LiverTox: Clinical and Research Information on Drug-Induced Liver Injury. 2018. Read the source