Category: Herbs & Ingredients
Ginkgo is sold as a memory herb, yet the largest, longest, best-run trial ever conducted on it found nothing. Here is the honest evidence for cognition and circulation, and the real bleeding and seizure cautions the marketing skips.
Ginkgo biloba has a problem most herbs would envy and none would want: it has been tested properly, and the biggest, best test came back negative. The tree itself is a marvel, a living fossil whose lineage predates most flowering plants, and its leaf extract has been one of the world's top-selling memory supplements for decades. But in the largest, longest and most rigorous trial ever run on it, more than 3,000 older adults took ginkgo or placebo for around six years, and the ginkgo group did not develop less dementia. If anything, numerically, they developed slightly more.
That single result reorganises everything an honest guide can say. It does not make ginkgo dangerous or worthless, but it does mean the confident memory claims on the packaging have outrun, and in the key trial contradicted, the evidence. So this guide leads with the negative trial rather than hiding it, grades ginkgo as the limited-to-insufficient herb it honestly is, and then spends real time on the thing that genuinely matters for your safety: a rare but serious bleeding risk that deserves more attention than the memory hype ever gets.
Ginkgo biloba is the leaf extract of the maidenhair tree, Ginkgo biloba, the sole surviving species of an ancient plant lineage. The part used in supplements is the leaf, and its activity is attributed to two groups of compounds: flavonoid glycosides, which are antioxidant plant pigments, and terpene lactones, namely the ginkgolides and bilobalide, which are proposed to affect blood flow and platelet activity. Better products are standardised, often to around 24% flavones, and the specific extract behind most good trials is known as EGb 761.
The honest, self-contained summary: ginkgo's most confident claim, protecting memory, is contradicted by the largest trial, which found no reduction in dementia over six years. For treating existing cognitive impairment the evidence is inconsistent and unreliable, and for tinnitus it is low certainty and probably no better than placebo. Europe registers a well-established use for cognitive impairment and mild dementia with the standardised extract, plus a traditional-use claim for minor circulatory symptoms, but both are more generous than the trial picture. The overall grade is limited to insufficient. It is well tolerated and easy on the liver, but the real safety story is a rare, serious bleeding risk, a seizure caution in epilepsy, specific drug interactions, and a firm avoid in pregnancy.
Ginkgo's proposed mechanisms are plausible on paper, which is part of why it took a very large trial to settle the question. Its flavonoids are antioxidants, and its terpene lactones are thought to improve blood flow, in part by acting on platelet-activating factor, a signal involved in clotting and inflammation. That combination gave rise to the long-standing idea that ginkgo boosts circulation to the brain and thereby sharpens thinking.
It is a neat theory, and it illustrates a recurring lesson in this field: a believable mechanism is a reason to test a herb, not evidence that it works. The same platelet-related activity that underpins the circulation story is also the basis of ginkgo's bleeding caution, so the mechanism cuts both ways. When the mechanism was finally put to a definitive clinical test for its headline use, preventing cognitive decline, it did not deliver, which is exactly why outcomes trials matter more than mechanistic promise.
Graded against the best available trials, ginkgo is one of the clearest examples of reputation outrunning proof.
Preventing dementia is the headline, and it is negative. The Ginkgo Evaluation of Memory study randomised 3,069 community-dwelling adults aged 75 and over to 120mg of ginkgo twice daily or placebo, and followed them for a median of about six years. Ginkgo did not reduce the overall rate of dementia or of Alzheimer's disease, and the hazard ratios slightly, though not significantly, favoured placebo [1]. This is the largest, longest and best-powered ginkgo trial ever conducted, and it found nothing. No memory claim can be made confidently over the top of that result.
Treating existing cognitive impairment is inconsistent. A Cochrane review of 36 trials concluded that the evidence that ginkgo has a predictable and clinically significant benefit for people with dementia or cognitive impairment is inconsistent and unreliable, while noting ginkgo appeared safe with no excess side effects versus placebo [2]. Inconsistent and unreliable is a fair summary of the whole cognition case.
Tinnitus is low certainty and probably null. A Cochrane review of 12 studies found ginkgo may have little to no effect on tinnitus severity compared with placebo, at low to very low certainty [3]. The US National Center for Complementary and Integrative Health lists ginkgo plainly as not helpful for tinnitus.
Circulation is traditional-use, not proven. Europe registers a traditional-use claim for the heaviness of legs and the sensation of cold hands and feet associated with minor circulatory disorders, but this rests on long-standing use and applies only after a doctor has excluded a serious cause. It is a modest, hedged claim, not a demonstration of efficacy.
What the evidence does not support is ginkgo as a memory booster for healthy people, a dementia preventive, or a reliable tinnitus remedy. The regulator's own overall position is that there is no conclusive evidence ginkgo is helpful for any health condition, with only limited hints of possible modest benefit in a few areas.
Ginkgo suits an adult who wants to try a standardised extract with realistic, limited expectations, and who has first ruled out the safety issues below.
Use a standardised extract. The trials used standardised leaf extract, so if you try ginkgo, choose a standardised product rather than generic powder, and never the seeds.
The dose: the standardised-extract studies used 120 to 240mg a day for at least eight weeks, with a review at around three months if nothing has improved. Ginkgo is not intended for anyone under 18.
Set honest expectations. Given that the best trial was negative, the sensible framing is a low-confidence trial of your own, stopped promptly if you notice no benefit, rather than an open-ended commitment based on reputation.
What to check first: your medication list, any bleeding tendency, any history of seizures, and pregnancy status. For ginkgo these are not footnotes; they are the main event.
Ginkgo is generally well tolerated in healthy adults and is not linked to liver injury, with the commonest side effects being headache, dizziness and mild digestive upset. But its safety profile has real edges, all drawn from published guidance and case reports rather than any product-specific claim.
Bleeding is the caution that matters. The picture is genuinely mixed and the responsible reading is caution. Controlled trials pooled together do not show a population-level rise in bleeding [5], yet the published literature contains real case reports of serious spontaneous bleeds, including bleeding within the skull [4]. Europe's monograph formally lists bleeding of individual organs as a possible effect. The way to hold both facts is this: the average person does not bleed more, but the rare idiosyncratic bleed can be catastrophic, and an unproven cognitive benefit is not worth that gamble for higher-risk people.
Surgery. Because ginkgo may increase susceptibility to bleeding, stop it as a precaution three to four days before any planned operation, and tell your surgical team.
Blood thinners and bleeding disorders. Anyone taking an anticoagulant or antiplatelet drug, or with a pathologically increased bleeding tendency, should use ginkgo only after consulting a doctor.
Epilepsy. Europe's monograph carries a genuine seizure caution, stating that in people with epilepsy, further seizures promoted by ginkgo cannot be excluded. If you have a seizure disorder, avoid ginkgo unless a specialist agrees otherwise.
Specific drug interactions. Beyond blood thinners, ginkgo may interact with certain medications, including the HIV medicine efavirenz, where combined use is not recommended, so a medication review is wise.
Pregnancy, breastfeeding and seeds. Ginkgo is contraindicated in pregnancy under its well-established use and is not recommended while breastfeeding. Separately, raw and roasted ginkgo seeds are toxic and have caused serious harm; only the standardised leaf extract belongs in a supplement, never the seeds.
Pregnant, breastfeeding, or on medication? Check with a healthcare professional first.
Ginkgo is the herb we point to when explaining why we lead with negative evidence instead of burying it. It would be easy to sell ginkgo the way most of the market does, with a confident line about memory and circulation and a quiet silence about the trial that failed. We would rather tell you that the definitive study came back empty, that the cognition evidence is officially inconsistent and unreliable, and that the most important thing about ginkgo is not its memory claim but its bleeding caution. A brand that hides the negative trial is a brand that will hide the next inconvenient fact too.
If you are weighing ginkgo against blood-thinning medication, read the ginkgo and blood thinners piece before starting anything, and for a herb often paired with it, the ginseng complete guide sets out its own honest evidence. For a cognition-focused botanical with its own carefully graded record, the bacopa memory herb piece is the better next read than reaching reflexively for ginkgo.
Ginkgo deserves a closing that respects the tree and the data at once. It is an extraordinary plant with an ordinary, and in its headline use disappointing, clinical record: a standardised extract whose biggest trial found no memory protection, whose cognition evidence the specialists call unreliable, and whose most consequential feature is a rare but serious bleeding risk. If you try it, do so with modest hopes, a standardised product, and a clear-eyed check that you are not in one of the groups it can harm. That is the honest offer, and with ginkgo, honesty means saying the negative trial out loud rather than hoping you never read it.
1. DeKosky ST, et al. (2008). Ginkgo biloba for prevention of dementia: a randomised controlled trial (the Ginkgo Evaluation of Memory study). JAMA. PMID 19017911. 2. Birks J, et al. (2009). Ginkgo biloba for cognitive impairment and dementia. Cochrane Database of Systematic Reviews. PMID 19160216. 3. Sereda M, et al. (2022). Ginkgo biloba for tinnitus. Cochrane Database of Systematic Reviews. PMID 36383762. 4. Bent S, et al. (2005). Spontaneous bleeding associated with ginkgo biloba: a case report and systematic review of the literature. Journal of General Internal Medicine. PMID 16050865. 5. Kellermann, et al. (2011). Is there a risk of bleeding associated with standardised ginkgo biloba extract therapy? A systematic review and meta-analysis. Pharmacotherapy. PMID 21923430.
The honest answer is that the strongest evidence says no, at least for preventing decline. The largest and longest trial, following more than 3,000 older adults for about six years, found ginkgo did not reduce the rate of dementia or Alzheimer's disease. Any benefit for everyday memory in healthy people is unproven.
Because reputation and older, smaller studies built an expectation that the largest, most rigorous trial then failed to confirm. This is a common pattern: early promise, then a big well-run study that deflates it. Honest grading follows the best evidence, which for ginkgo is limited to insufficient.
EGb 761 is the specific standardised ginkgo leaf extract used in most of the better-quality trials, standardised to fixed proportions of flavonoid glycosides and terpene lactones. A random ginkgo product is not automatically equivalent to the extract that was actually studied.
The evidence says probably not. A Cochrane review found ginkgo may have little to no effect on tinnitus severity compared with placebo, at low certainty, and the US National Center for Complementary and Integrative Health lists it as not helpful for tinnitus.
Europe registers a traditional-use claim for the heaviness of legs and the sensation of cold hands and feet linked to minor circulatory problems, but this is based on long-standing use rather than strong trial evidence, and only after a doctor has excluded a serious cause.
The standardised-extract trials used a daily dose of 120 to 240mg for at least eight weeks. Europe advises reviewing after about three months if there is no improvement. It is not intended for anyone under 18.
For most healthy adults, standardised ginkgo leaf extract is generally well tolerated and is not linked to liver injury. The main concerns are a rare bleeding risk, a seizure caution in epilepsy, and specific drug interactions, all covered below.
This is the caution to take seriously. Although controlled trials do not show a population-level rise in bleeding, real case reports describe serious spontaneous bleeds, including in the brain. With the stakes that high, ginkgo should only be combined with anticoagulant or antiplatelet medication after speaking to your prescriber. This comes from the published literature, not from any product claim.
Europe's monograph carries a genuine seizure caution, stating that in people with epilepsy, further seizures promoted by ginkgo cannot be excluded. If you have a seizure disorder, this is a reason to avoid ginkgo unless a specialist agrees otherwise.
Yes. Because ginkgo may increase susceptibility to bleeding, guidance is to stop it as a precaution three to four days before surgery. Tell your surgical team you have been taking it.
No, and this matters. Raw and roasted ginkgo seeds are toxic and have caused serious harm, whereas the standardised leaf extract sold as a supplement is a different preparation. Never eat ginkgo seeds as a home remedy.