Supplement · graded by use, three uses
Ginkgo Biloba (EGb 761)
Cognitive symptoms in mild to moderate dementia.
Ginkgo is an extract made from the leaf of the ginkgo tree. The studies used one standard extract at 240 mg a day. In people with mild to moderate dementia, pooled studies of more than 2,500 patients found better thinking scores over about six months. Other reviews disagreed, and most positive studies had ties to the maker. For mild thinking problems, only one study exists. Case reports link it to bleeding with blood thinners, though tests with warfarin found little effect. It may weaken seizure drugs.
The B belongs to one of the three uses graded on this record. The other two carry their own letters.
As studied, and the trials
Tan et al. (J Alzheimers Dis, 2015; PMID 25114079) and Gauthier and Schlaefke (Clin Interv Aging, 2014; PMID 25506211) found cognition improved; Riepe et al. (World J Biol Psychiatry, 2025; PMID 39895346) agrees in mild dementia. Birks and Grimley Evans (Cochrane, 2009; PMID 19160216) and Thancharoen et al. (Drugs Aging, 2019; PMID 30937879) dissent. Replicated The same finding reproduced by a separate research group. A result seen once can be chance or bias; a replicated one is harder to explain away. trials with dissenting pools fit B, not A. The replicating trials are largely manufacturer-affiliated, which is disclosed here.
As studied, and the trials
Gavrilova et al. (Int J Geriatr Psychiatry, 2014; PMID 24633934): neuropsychiatric symptoms and a trail-making test improved against placebo An inactive substitute made to look like the real treatment, so participants cannot tell which one they received.. One trial with manufacturer-affiliated authors, not independently repeated: C.
As studied, and the trials
GEM (DeKosky et al., JAMA, 2008; PMID 19017911): 3,069 adults aged 75 and over, 120 mg twice a day for a median 6.1 years; dementia 3.3 against 2.9 per 100 person-years (hazard ratio How often an event happened in the treated group compared with the control group over the same period. A ratio of 0.80 means 20 percent fewer events. 1.12, 0.94 to 1.33), and no effect on progression in those who entered with mild cognitive impairment. GuidAge (Vellas et al., Lancet Neurol, 2012; PMID 22959217): 2,854 adults aged 70 and over with memory complaints, the same dose of EGb 761 for 5 years; probable Alzheimer's disease 1.2 against 1.4 per 100 person-years (hazard ratio 0.84, 0.60 to 1.18), in a trial funded by the extract's maker. Two large trials, neither preventing dementia: tested, no benefit. This does not grade symptoms in people who already have dementia, which are graded on their own rows.
Read the grade that matches your goal. How EBL grades by use →
- Dose studied
- 240 mg/day (EGb 761)
- Form graded
- Standardized EGb 761 leaf extract
- Evidence behind the B
- PMID 25114079 →PMID 25506211 →PMID 39895346 →PMID 19160216 →
A letter describes the human evidence, not the product, the manufacturer, or the ingredient's worth.
See the full evidence record →Ginkgo biloba (EGb 761) for cognitive symptoms in mild to moderate dementia, graded on human trial evidence.
As studied: dose, form, population and duration for every use
A grade describes the intervention that was tested. Where the product on the shelf differs from it, this says so.
How this letter has changed · latest: October 2026 · Added
Two large randomized trials of the standardized extract, GEM (DeKosky et al., JAMA, 2008; PMID 19017911) in 3,069 adults aged 75 and over and GuidAge (Vellas et al., Lancet Neurol, 2012; PMID 22959217) in 2,854 adults with memory complaints, did not prevent dementia or Alzheimer's disease. Prevention now has its own row at D, tested and no benefit. The rows on symptoms in people who already have dementia, and the record's top letter, do not change.PMID 19017911 PMID 22959217
The record called both of its citations meta-analyses A study that pools the numbers from several trials into one combined estimate. of EGb 761 covering dementia and mild cognitive impairment. Liao et al. (PMID 32658034) did not name EGb 761, and Riepe et al. (PMID 39895346) covered mild dementia only. A PubMed search added Tan et al. (J Alzheimers Dis, 2015; PMID 25114079) and Gauthier and Schlaefke (Clin Interv Aging, 2014; PMID 25506211) for dementia, the dissenting Birks and Grimley Evans (Cochrane, 2009; PMID 19160216) and Thancharoen et al. (Drugs Aging, 2019; PMID 30937879), and the one blinded trial in mild cognitive impairment, Gavrilova et al. (Int J Geriatr Psychiatry, 2014; PMID 24633934), which is graded C on its own.PMID 32658034 PMID 39895346 PMID 25114079 PMID 25506211 PMID 19160216 PMID 30937879 PMID 24633934
Every change to a letter on this site is dated and published. The full evidence log →
Commercial disclosure
EBL earns no commission on any supplement it grades. A purchase made through the dispensary link below pays EBL nothing, the practitioner discount goes to you in full. The dispensary is Fullscript, a practitioner dispensary, and many of the professional-grade brands in it are sold mainly through practitioners rather than in health food stores. Grades are assigned by published rubric before any commercial relationship exists and are never influenced by one. Full disclosure.
No dose is published here. This record carries no authored dose, because the cited trials do not converge on a single figure. Each trial cited on this page states the dose it used, and EBL has not turned those into a recommendation.
Standardized extracts were used in the cited trials; product quality varies.
The cited reviews pool trials of the extract and do not report when doses were taken (PMID 32658034, PMID 39895346).
Checked against the published summaries; the full papers could not be read. What the cited trials did, not a schedule. Every supplement, by trial timing →
Seizure disorders. Ginkgo seeds, and to a much smaller degree leaf extracts, contain 4'-O-methylpyridoxine, which antagonizes vitamin B6. Surgery: two weeks of separation is the usual clinical margin. Pregnancy.
Anticoagulants and antiplatelets, including warfarin, aspirin and clopidogrel. Ginkgolides antagonize platelet-activating factor, which is the mechanism behind the warning, and published case reports exist. When it was tested, randomized trials found no clinically meaningful effect on INR, on warfarin pharmacokinetics, or on warfarin dose requirement, so the honest position is a monitoring caution rather than a prohibition. Ginkgo induces CYP2C19, which can lower blood levels of drugs cleared by that enzyme. Anticonvulsant effectiveness may be reduced.
The ginkgo warfarin interaction was tested and did not appear. Randomized trials of ginkgo in warfarin-treated participants found no clinically meaningful effect on INR, on warfarin pharmacokinetics, or on dose requirement, so the case-report-based warning is stronger than what testing showed. Jiang et al. (British Journal of Clinical Pharmacology, 2005; PMID 15801937): randomized trial of the effect of ginkgo and ginger on the pharmacokinetics and pharmacodynamics of warfarin in healthy subjects. Engelsen et al. (Thrombosis and Haemostasis, 2002; PMID 12083489): randomized double-blind placebo-controlled crossover A trial design in which each participant receives both the treatment and the comparison, one after the other, and serves as their own control. of Ginkgo biloba on warfarin dosage in stable long-term warfarin outpatients.
A validated cognitive assessment (MoCA or ADAS-Cog) at baseline and after 12 weeks to measure the memory and attention outcomes studied in the EGb 761 trials. Subjective cognitive complaints using a standardized self-report scale alongside the objective test. EGb 761 is the specific extract studied: results with other ginkgo preparations are not established.
Primary evidenceSeven papers, each linked to its PubMed record.
Tan et al. (Journal of Alzheimer's Disease, 2015; PMID 25114079): 9 randomized trials of EGb 761, 2,561 patients, 22 to 26 weeks; cognition, daily activities and clinician global impression favored EGb 761, mainly at 240 mg a day. Gauthier and Schlaefke (Clinical Interventions in Aging, 2014; PMID 25506211): 7 trials, cognition improved. Riepe et al. (World Journal of Biological Psychiatry, 2025; PMID 39895346): pooled subgroup analysis of 4 placebo-controlled trials in 782 patients with mild dementia at 240 mg a day. Liao et al. (Clinical Neuropharmacology, 2020; PMID 32658034): 7 studies, 939 patients with Alzheimer disease, of Ginkgo biloba preparations not named as EGb 761. Not every pool agrees: Birks and Grimley Evans (Cochrane Database of Systematic Reviews A review that searches for every study on a question using a method set in advance, then assesses them together., 2009; PMID 19160216) judged the evidence inconsistent, and Thancharoen et al. (Drugs and Aging, 2019; PMID 30937879) found no benefit over placebo in mild to moderate Alzheimer disease. Several replicating trials and two of the pooled analyses have authors from the manufacturer. Gavrilova et al. (International Journal of Geriatric Psychiatry, 2014; PMID 24633934) is the one blinded trial in mild cognitive impairment.
Educational use only · Not medical advice · Starting, stopping or changing any supplement or medication is a decision for your physician
These statements have not been evaluated by the Food and Drug Administration. This product is not intended to diagnose, treat, cure, or prevent any disease. The grade on this page describes the strength of published human trial evidence in the populations studied. It is not a recommendation for any individual and is not medical advice. Starting, stopping or changing any supplement or medication is a decision for your physician.
Who responds: this record's trial reading reports no subgroup that identifies who responds.
This page carries a letter. EBL takes no commission, referral fee or payment of any kind on anything it grades. No letter has been bought and none is for sale.
How this is enforced
That is not a promise, it is a build condition. Two checks run on every deploy: one refuses to publish if a commission is claimed on anything graded, the other refuses if a page names a commercial relationship that is not declared in the register. If either fails, this page does not go live. You can read how the checks work, and what they do not cover, on the evidence integrity page.
How EBL searched for the trials (1 search)
The search that was run, what it returned, and what was left out and why, so the absence of evidence can be checked as well as its presence.
("EGb 761"[tiab] OR EGb761[tiab]) AND (dementia[tiab] OR "mild cognitive impairment"[tiab] OR "cognitive impairment"[tiab]) AND (meta-analysis[pt] OR randomized controlled trial[pt])Grades move when new trials publish. A free monthly email lists every grade EBL moved, with the reason and the trial behind it: the sign-up is on the Evidence Log →