- Ginkgo does not prevent dementia. The US Ginkgo Evaluation of Memory (GEM) trial followed 3,069 people aged 75 or older for a median of 6.1 years and found no benefit: hazard ratio 1.12 (95% CI 0.94 to 1.33) for all-cause dementia, where 1.0 means no difference (DeKosky et al., JAMA 2008). The US National Center for Complementary and Integrative Health (NCCIH) says ginkgo has not been shown to prevent or slow dementia (NCCIH).
- A second, manufacturer-funded trial agrees. GuidAge randomised 2,854 French adults aged 70 or older with memory complaints and found hazard ratio 0.84 (95% CI 0.60 to 1.18) for Alzheimer's disease over 5 years, which is not a significant difference (Vellas et al., Lancet Neurology 2012). The funder was the drug company Ipsen.
- For people who already have dementia, the Cochrane review judged the evidence for ginkgo "inconsistent and unreliable". It found 36 trials, most small and short, and a pattern in which many of the larger trials were funded by the extract maker (Birks & Grimley Evans, Cochrane 2009; full text). The European Medicines Agency (EMA) nonetheless lists a standardised dry extract for age-related cognitive impairment and mild dementia (EMA monograph).
- Bacopa has more modest and less certain evidence. Two small meta-analyses found improved memory recall or attention in healthy adults over 12 weeks (Pase 2012; Kongkeaw 2014), but a 2023 review of herbal and nutritional trials in older adults, 6 of 21 of them on bacopa, rated 14 of the 21 at high risk of bias (Cave et al.), while the Kongkeaw meta-analysis rated its own included trials low risk. Citicoline has an older Cochrane review with limited evidence in elderly people with chronic cerebral disorders (Cochrane 2005).
- Ginkgo affects bleeding and several drugs. The EMA says people with bleeding tendency or on anticoagulants or antiplatelet drugs should use it only after consulting a doctor, and that it should be stopped 3 to 4 days before surgery (EMA monograph). GEM found no excess total bleeding but a non-significant doubling of haemorrhagic stroke (16 vs 8 cases) (GEM full text).
- Money trail: GEM was publicly funded (NCCIH and the National Institute on Aging) but the extract maker, Schwabe, donated the tablets and placebos (GEM full text). GuidAge was paid for by Ipsen (PubMed record). The Cochrane review notes that Schwabe funded several earlier trials and that one trial, DIGGER, was independent of the maker (Cochrane full text). In January 2017 the US Federal Trade Commission (FTC) and New York State charged the makers of Prevagen with deceptive memory claims (FTC); the FTC reported a win in December 2024 (FTC statement).
- Evidence grade: Moderate that ginkgo does not prevent dementia. Weak for ginkgo in existing dementia and for bacopa in memory. Insufficient for healthy-adult memory boosting, citicoline as a supplement, phosphatidylserine and Prevagen. Risk for bleeding and drug interactions with ginkgo.
Independent evidence review · Supplement
The best independent evidence says ginkgo does not stop memory loss turning into dementia, and the case for bacopa and other "memory" supplements rests on small, short trials of mixed quality. Two large dementia-prevention trials, one publicly funded and one funded by a drug company, both came out negative. Benefits in people who already have dementia are inconsistent, and the larger positive trials often came from the maker of the leading extract. Ginkgo can raise bleeding risk and interact with several medicines. Our confidence: moderate-to-high that ginkgo fails for prevention; low for every other memory claim covered here.
Table of contents
- Evidence summary
- What these supplements are, and their legal status
- How they are thought to work
- What works and what does not
- Benefits by claim
- Risks and side effects
- Interactions
- Who should avoid ginkgo
- Doses studied and product quality
- Follow the money: who makes it and who funded the evidence
- Related research
- Frequently asked questions
- Sources and funding notes
Evidence summary
| Claim | Evidence | Source | Funding / conflict | Strength |
|---|---|---|---|---|
| Ginkgo prevents dementia or Alzheimer's disease | GEM: 3,069 people aged 75+, 120 mg twice daily, median 6.1 years. Dementia rate 3.3 vs 2.9 per 100 person-years (ginkgo vs placebo); HR 1.12 (95% CI 0.94 to 1.33). GuidAge: 2,854 people, 5 years, HR 0.84 (0.60 to 1.18) for Alzheimer's; risk not proportional over time. | DeKosky 2008; Vellas 2012 | GEM: NCCIH and National Institute on Aging; extract and placebo donated by Schwabe. GuidAge: funded by Ipsen. | Moderate (against the claim) |
| Ginkgo helps people who already have dementia or cognitive impairment | Cochrane: 36 trials, nine of about six months (2,016 people). Results inconsistent for cognition, daily living and mood; no consistent benefit in the Alzheimer's subgroup (925 people, nine trials). EMA lists a dry extract for age-associated cognitive impairment and mild dementia. | Birks 2009; EMA monograph | Many included trials funded by Schwabe, the maker of EGb 761 (per the Cochrane text). | Weak |
| Ginkgo or other supplements boost memory in healthy adults | NCCIH: research in healthy people is generally low quality. A review of nine popular ingredients found inconsistent, imprecise and often flawed studies. A 2023 review: 14 of 21 herbal and nutritional trials had high risk of bias. | NCCIH; Crawford 2021; Cave 2023 | NCCIH: US government. Crawford: funding not verified. Cave: NHMRC grants and a PhD scholarship from FIT-BioCeuticals, a supplement company. | Insufficient |
| Bacopa improves memory | Systematic review of six 12-week trials (300 to 450 mg extract daily): improved memory free recall on 9 of 17 tests, little effect elsewhere. Meta-analysis of nine trials (518 people): faster Trail B and choice reaction time. | Pase 2012; Kongkeaw 2014 | Funding statements could not be accessed for either review. | Weak |
| Citicoline (CDP-choline) improves memory | Cochrane: 14 studies in elderly people with chronic cerebral disorders; some evidence of benefit on memory and behaviour in the short to medium term, none on attention. Studies were short and varied. | Cochrane 2005 | Funders of included trials not checked. | Weak |
| Phosphatidylserine reduces dementia risk | The FDA has issued qualified health claim letters (2003, 2004) on phosphatidylserine and cognitive function and dementia. We could not open the letter text, and found no independent trial review in this session. | FDA qualified health claims | US regulator | Insufficient |
| Prevagen (apoaequorin) improves memory | FTC and New York State alleged the company's own study failed to show better results than placebo on any cognitive measure. A court ruled for the FTC in 2024. | FTC 2017; FTC 2024 | US government (enforcement); product maker is Quincy Bioscience. | Insufficient |
| Ginkgo increases bleeding | GEM: total bleeding 138 vs 140 events (HR 0.97; 0.77 to 1.23); haemorrhagic stroke 16 vs 8 (HR 1.97; 0.84 to 4.61, not significant). EMA: bleeding of eye, nose, brain and gut reported, frequency unknown. A review links ginkgo with bleeding risk in people on anticoagulants but calls overall bleeding evidence conflicting. | GEM full text; EMA; Hatfield 2022 | As above; Hatfield: no funding statement found in the record. | Risk |
| Ginkgo extract caused tumours in animals | US National Toxicology Program: "some evidence" in rats and "clear evidence" of liver tumours in mice at the doses tested by gavage. Human relevance not established. | NTP TR-578 | US government (NIH) | Risk (animal data) |
Independent evidence and credibility scorecard
Independence tiers: 1 = regulator or government body; 2 = academic or Cochrane team with no declared commercial funding for the work; 3 = independent authors analysing mostly manufacturer-run trials, or minor declared industry ties; 4 = manufacturer-funded or manufacturer-authored. Credibility runs from A (highest) to D.
| Source | Who funds it | Country | Independence tier (1–4) | Credibility (A–D) | Why it would tell the truth / residual bias |
|---|---|---|---|---|---|
| NCCIH ginkgo fact sheet | US federal government (part of the National Institutes of Health) | US | 1 | A− | Public-health mandate; no product to sell. NCCIH also co-funded GEM, so it reports on its own investment. |
| EMA ginkgo monograph and assessment report | EU agency (public body) | EU | 1 | A− | Legal mandate to protect patients. The "well-established use" status rests on the published trial record, which we did not audit trial by trial for funding. |
| FTC Prevagen complaint, case page and 2024 statement | US federal government | US | 1 | A− | Consumer-protection mandate. A complaint is an allegation; the 2024 statement reports the outcome. We have not seen the court's full reasoning. |
| FDA warning letter (Peak Nootropics), with FDA Alzheimer's products page | US federal government | US | 1 | A− | Regulatory mandate. Letters address marketing claims, not whether bacopa works; they reflect the seller's website at one date. |
| NTP ginkgo study (TR-578) | US federal government (NIH) | US | 1 | A | Standardised animal testing with no product to sell. Animal data and high doses; relevance to supplement users is uncertain. |
| GEM trial (DeKosky 2008) | NCCIH, National Institute on Aging and other NIH institutes; Schwabe donated product | US | 2 | A− | Large, long, pre-registered and publicly run; the sponsor helped design and analyse it, and the maker had no such role per the paper. Donated product is an indirect tie. Lead author declared pharma ties on unrelated drugs. |
| Cave et al. 2023 review | Australia's National Health and Medical Research Council; PhD scholarship from FIT-BioCeuticals (a supplement company) | Australia | 2 | B | Systematic and transparent about bias. Authors say funders played no role; the scholarship from a supplement company is a minor tie. |
| Cochrane review of citicoline (2005) | Cochrane review group (funders not verified) | UK (Cochrane) | 2 | B− | Independent method, but searches ended in 2004 and trials were short and varied. Trial funders not checked. |
| Pase et al. 2012 bacopa review | Funding statement not accessed | Not verified | 2 | C+ | Provisional tier. Only six trials, three different extracts. We could not verify authors' ties to bacopa trial sponsors. |
| Kongkeaw et al. 2014 bacopa meta-analysis | Funding statement not accessed | Not verified | 2 | C+ | Provisional tier. Nine trials, 518 people; the abstract's units for the two main results look inconsistent, so we quote it cautiously. |
| Birks & Grimley Evans, Cochrane 2009 | Cochrane Dementia group; Alzheimer's Society supported an earlier version; no author conflicts declared in the record | UK | 3 | B | Reputation depends on rigour, and it published an unfavourable conclusion. Many of the trials it pooled were funded by the extract maker, and searches ended in 2007. |
| GuidAge (Vellas 2012) | Ipsen (drug company) | France (trial setting) | 4 | B− | Large, pre-registered and published a null result against its funder's interest. Funder-run trials can still shape outcomes and analyses. Direct tie to the product. |
Note on clustering: four of the twelve graded sources are US federal bodies (NCCIH, FTC, FDA, NTP), and a fifth, GEM, is US-funded, so the verdicts partly reflect a US lens. The EMA and Cochrane add European views. On dementia prevention, the publicly funded GEM trial and the Ipsen-funded GuidAge trial, whose interests differ, reach the same null result. That convergence between opposing interests is the strongest signal in this article.
What these supplements are, and their legal status
Ginkgo biloba is a tree whose leaf extract is sold as a memory supplement and, in parts of Europe, as a licensed herbal medicine. The best-studied product is a standardised extract called EGb 761, tested in GEM, GuidAge and most longer Cochrane-reviewed trials (Cochrane; GuidAge). The EMA monograph describes a dry extract (35 to 67 parts of leaf per part of extract, made with 60% acetone) (EMA monograph).
Bacopa monnieri is an Ayurvedic herb. The trials we reviewed used three different extracts at 300 to 450 mg a day (Pase 2012). Citicoline (CDP-choline) is a form of choline prescribed as a drug in several European countries for cognitive impairment, according to the Cochrane review (Cochrane 2005). Phosphatidylserine is a fat molecule sold for memory. Prevagen contains apoaequorin, a jellyfish-derived protein (the FTC describes the active ingredient as a protein derived from jellyfish) (FTC).
Legal status
- EU: Ginkgo leaf dry extract has an EU herbal monograph for "well-established use" in age-associated cognitive impairment and mild dementia, and powdered leaf has a traditional-use indication for heavy legs and cold hands and feet. The monograph was adopted on 28 January 2015 (EMA monograph; EMA product page).
- US: The FDA says it has sent warning letters to companies selling products as unapproved drugs for Alzheimer's disease. Bacopa is named in one of them; ginkgo is not named on that page (FDA). A February 2019 warning letter said the seller's bacopa claims (including preventing Alzheimer's) made it an unapproved new drug (FDA warning letter). The FDA has also issued qualified health claim letters on phosphatidylserine, dated 2003 and 2004 (FDA).
How they are thought to work
For ginkgo, the EMA says the exact mechanism is not known. Human pharmacology data it cites show increased EEG vigilance in older people, reduced blood viscosity, improved blood flow in some brain areas in healthy men aged 60 to 70, reduced platelet clumping and widened forearm blood vessels (EMA monograph). Those are effects on measurements, not proof of better memory or less dementia: the large trials above tested the outcome directly and found none (GEM).
For bacopa, citicoline, phosphatidylserine and apoaequorin, we did not open a source that establishes a mechanism in people, so we do not describe one. The FTC complaint alleges the Prevagen maker had no evidence that swallowed apoaequorin reaches the brain; that is an allegation as stated in the complaint summary, not a finding we can independently confirm (FTC).
What works and what does not
| Use | Verdict | Why |
|---|---|---|
| Ginkgo to prevent dementia | Does not work | Two large trials, one public and one drug-company funded, both null (GEM; GuidAge). |
| Ginkgo in established dementia | Mixed | Cochrane found inconsistent, unreliable results; EMA accepts well-established use (Cochrane; EMA). |
| Bacopa for memory recall in healthy adults | Mixed | Small effects on some tests; few trials and risk-of-bias ratings differ between reviews (Pase; Kongkeaw; Cave). |
| Citicoline in elderly people with cerebrovascular disease | Mixed | Some memory and behaviour benefit in short trials (Cochrane). |
| Phosphatidylserine, Prevagen and "memory blends" | Insufficient | No independent trial review opened; FTC succeeded against Prevagen's claims (FTC). |
Benefits by claim
Preventing dementia or Alzheimer's disease
GEM enrolled 3,069 community volunteers aged 75 or older, 2,587 with normal thinking and 482 with mild cognitive impairment, at five US academic centres. Of 523 who developed dementia, 277 took ginkgo and 246 placebo. The hazard ratio was 1.12 (95% CI 0.94 to 1.33) for all dementia and 1.16 (0.97 to 1.39) for Alzheimer's disease. In plain English, the ginkgo group did not do better; if anything the numbers leaned slightly the other way, but not by enough to be sure (DeKosky 2008).
GuidAge randomised 2,854 adults aged 70 or older with memory complaints to 120 mg of EGb 761 twice daily or placebo. By 5 years, 61 ginkgo and 73 placebo participants had been diagnosed with probable Alzheimer's (HR 0.84, 95% CI 0.60 to 1.18, p=0.306). The authors noted the risk was not proportional over time, and concluded ginkgo did not reduce progression to Alzheimer's (Vellas 2012). The funder was Ipsen. A company-funded trial that fails to favour its own product type is notable, and it matches the publicly funded GEM result.
Treating existing dementia
The Cochrane review found 36 trials, mostly small and under three months. Nine trials lasted about six months and included 2,016 people. Most tested EGb 761. Of the four most recent trials with results at the time, three found no difference from placebo and one found very large benefits (Cochrane abstract). The full text shows that several trials acknowledged funding by Schwabe, the extract maker, and that DIGGER was independent of it. In the six-month trials, the pooled cognition difference was 0.38 standard deviations for any dose, but it fell to −0.04 (95% CI −0.14 to 0.06; negative favours ginkgo) when two trials, Napryeyenko 2005 and Mazza 2006, were left out; the review also reported very high heterogeneity on one scale (I² 97%) (Cochrane full text). That fragility is why the authors called the evidence unreliable.
A 2020 meta-analysis of seven studies (939 people) in Alzheimer's disease, including Chinese-language databases, reported a risk ratio of 1.98 (95% CI 1.52 to 2.59) for cognitive improvement with ginkgo preparations (Liao et al. 2020). We did not verify its funding, and Chinese trials of herbal products have often had weaker methods, so we weigh it below the Cochrane review. EMA's "well-established use" status supports the extract for mild dementia (EMA), but the agency itself says treatment should be reviewed if there is no improvement after 3 months.
Improving memory in healthy people
NCCIH says research on whether ginkgo improves memory in healthy people is generally low quality (NCCIH). A systematic review of nine popular "cognitive performance" ingredients, including ginkgo and bacopa, concluded that studies were inconsistent, imprecise and often flawed (Crawford et al. 2021). A 2023 review of 21 trials in older adults found 14 with high risk of bias and one with low risk; the low-risk trial tested a bacopa-containing combination (Cave et al., full text).
Bacopa
Pase and colleagues reviewed six trials of 12 weeks each in adults without dementia, using three extracts at 300 to 450 mg daily. Bacopa improved performance on 9 of 17 memory free recall tests, with little evidence in other areas (Pase 2012). Kongkeaw and colleagues meta-analysed nine trials (518 people, with 437 in the main analysis) and reported faster Trail B completion (−17.9, 95% CI −24.6 to −11.2) and shorter choice reaction time, concluding Bacopa may help speed of attention (Kongkeaw 2014). The abstract labels the Trail B unit as milliseconds, which looks like a labelling slip. The authors themselves called for a large head-to-head trial. Kongkeaw rated the included trials low risk of bias, whereas Cave and colleagues rated most of the herbal and nutritional trials they pooled (bacopa among them) high risk; a modest effect on recall and attention speed, from short trials, is as far as the evidence goes.
Citicoline and phosphatidylserine
The Cochrane review of citicoline pooled 14 studies in older people with chronic cerebral disorders. It found some evidence of benefit on memory and behaviour in the short to medium term, none on attention, and good tolerability (Cochrane 2005). Those were mostly short treatment courses in people with vascular disease, not healthy adults taking a supplement. For phosphatidylserine, the FDA has issued letters dated 24 February 2003, 13 May 2003 and 24 November 2004 on a qualified health claim for cognitive function and dementia (FDA). A qualified claim means the agency allowed limited wording, not that benefit is proven; we could not open the letters, so we do not quote them.
Prevagen
On 9 January 2017 the FTC and New York State filed a complaint against Quincy Bioscience Holding Company and related companies and two co-founders. They alleged false and unsubstantiated claims that Prevagen improves memory and is "clinically shown" to work, and that the company's double-blind study did not beat placebo on any cognitive measure. The complaint cited sales above $165 million (FTC). The FTC case page lists a Second Circuit summary order on 21 February 2019 and a Memorandum and Judgment on 18 November 2024 (FTC case page). In a statement of 10 December 2024 the FTC said the court ruled that Prevagen's marketing made deceptive claims and ordered the company to stop making the challenged claims (FTC statement). The statement does not give the order's full terms, and we have not seen the court's reasoning.
Risks and side effects
Ginkgo
The EMA lists headache as very common, and dizziness, diarrhoea, abdominal pain, nausea and vomiting as common. Allergic reactions, including skin rash and, rarely, allergic shock, can occur (EMA monograph). NCCIH lists dizziness, stomach upset and headache as the most common effects (NCCIH). In GEM, adverse effect profiles were similar in both groups (GEM). The Cochrane review found no significant difference in adverse events from placebo (Cochrane).
In GEM, total bleeding did not differ (138 on ginkgo vs 140 on placebo), but haemorrhagic stroke occurred in 16 vs 8 people (HR 1.97, 95% CI 0.84 to 4.61), which the authors judged non-significant but worth further study (GEM full text). GuidAge found no difference in haemorrhagic or cardiovascular events and a stroke risk ratio of 1.12 (0.77 to 1.63) (GuidAge).
Long-term animal data
The US National Toxicology Program gave ginkgo extract by gavage to rats and mice for two years. It found "some evidence" of carcinogenic activity in rats (thyroid) and "clear evidence" in mice (liver), plus liver, thyroid and nose lesions, and the extract was mutagenic in some bacteria (NTP TR-578). We did not open the dose levels, so we cannot say how they compare with supplement doses. No human cancer signal appeared in the trials we opened, but those trials were not designed to detect it.
Bacopa
In the trials reviewed by Cave and colleagues, one bacopa trial found significantly more side effects on treatment, including more frequent stools, nausea and abdominal cramps. Another reported digestive problems and flu-like symptoms in both groups, and one participant withdrew with a rash. The review noted that safety reporting was inconsistent across studies (Cave full text). We did not verify liver or long-term safety data for bacopa.
Interactions
- Anticoagulants and antiplatelets (warfarin, phenprocoumon, clopidogrel, aspirin, anti-inflammatory painkillers): the EMA says their effect may be influenced. Studies with warfarin did not indicate an interaction, but it advises monitoring when starting, changing or stopping ginkgo, or switching product (EMA monograph). A review of supplements and bleeding associates ginkgo with bleeding risk in people on anticoagulants, calls evidence of bleeding without anticoagulants conflicting, and suggests stopping non-essential supplements 2 weeks before surgery (Hatfield 2022).
- Dabigatran: a study with talinolol suggests ginkgo may inhibit gut P-glycoprotein, so caution is advised (EMA monograph).
- Nifedipine: one study found peak levels up to 100% higher in some people, with dizziness and worse hot flushes (EMA monograph).
- Efavirenz (HIV medicine): not recommended together, as ginkgo may lower its levels by inducing CYP3A4 (EMA monograph).
- Epilepsy medicines: new seizures promoted by ginkgo cannot be excluded in people with epilepsy (EMA monograph).
- Bacopa and others: we did not open a source on bacopa, citicoline or phosphatidylserine interactions, so we make no claim either way.
Our pages on anticoagulants and bleeding safety and supplements and heart medicines give wider context.
Who should avoid ginkgo
- Pregnancy: the EMA monograph contraindicates the dry extract in pregnancy because it may impair platelet aggregation and raise bleeding tendency (EMA monograph). NCCIH says ginkgo may be unsafe in pregnancy (NCCIH).
- Breastfeeding: the EMA says use is not recommended in the absence of data (EMA monograph); NCCIH says safety is not well known (NCCIH).
- Children and teenagers: no relevant indication, and use under 18 is not established (EMA monograph).
- People with bleeding disorders, on blood thinners, or facing surgery: only after a doctor's advice; stop 3 to 4 days before surgery per the EMA (EMA monograph).
- People with epilepsy or on efavirenz: see interactions above.
- Anyone with new memory loss: needs assessment first. Reversible causes exist, and supplements can delay care. The FDA warns that unproven Alzheimer's products can keep people from proper diagnosis and treatment (FDA).
Doses studied and product quality
Doses studied
GEM and GuidAge both used 120 mg of EGb 761 twice daily (GEM; GuidAge). The EMA monograph for the dry extract gives a single dose of 120 to 240 mg and a daily dose of 240 mg, a minimum course of 8 weeks and a review at 3 months; for powdered leaf under traditional use it gives a daily dose of 750 mg (EMA monograph). Bacopa trials used 300 to 450 mg of extract daily for 12 weeks (Pase 2012). We did not verify citicoline or phosphatidylserine doses. These are trial and label doses, not recommendations.
Product quality
The extract matters. The trials above used one specific, standardised, quality-controlled extract; a shop product may differ. The EMA assessment report notes the Ph. Eur. standard for refined dry extract limits ginkgolic acids to a maximum of 5 ppm (EMA assessment report). Crawford and colleagues warn that "not always knowing what is in a dietary supplement product" makes weighing risks and benefits difficult (Crawford 2021). A widely cited 2013 DNA-barcoding study of herbal products that included ginkgo products has since been retracted, so we do not rely on it (retracted BMC Medicine paper). We found no independent assay of ginkgo or bacopa shop products in this session. See our COA transparency audit for how supplement brands publish lab results.
Follow the money: who makes it and who funded the evidence
Who makes and sells it
Dr. Willmar Schwabe GmbH, based in Karlsruhe, Germany, supplied the EGb 761 tablets and placebo for GEM (GEM full text). The Cochrane review describes Schwabe as the manufacturer of the most widely used standardised extract and records that one of its employees submitted feedback to the review (Cochrane full text). Ipsen, a pharmaceutical company, funded GuidAge (PubMed record). Quincy Bioscience sells Prevagen (FTC). We could not trace the ownership of Schwabe, Ipsen or Quincy from the sources opened.
Who funded the trials
- GEM: grants from NCCIH (with the Office of Dietary Supplements and National Institute on Aging), plus other NIH institutes and several universities. Schwabe donated tablets and, according to the paper, had no role in design, analysis or writing, although three Schwabe staff are listed among the GEM study investigators. The lead author reported research grants and advisory roles with pharmaceutical firms (GEM full text).
- GuidAge: Ipsen (PubMed record).
- Earlier dementia-treatment trials: Cochrane says Kanowski 1996, Le Bars 1997 and Schneider 2005 acknowledge funding from Schwabe, while DIGGER was one of the few independent trials (Cochrane full text).
- Bacopa trials: we could not access the funding statements of the two meta-analyses and did not identify who funded the individual trials.
Documented integrity events
- FTC and New York State v. Quincy Bioscience: complaint filed 9 January 2017; Memorandum and Judgment dated 18 November 2024 and an order dated 6 December 2024; FTC statement of 10 December 2024 reporting a win (FTC; FTC case page; FTC statement).
- FDA warning letter dated 4 February 2019 to Peak Nootropics LLC of Houston, Texas, covering 13 products including bacopa, for disease claims including Alzheimer's prevention (FDA).
- FDA action on unproven Alzheimer's products: 12 warning letters and 5 online advisory letters covering more than 58 products (FDA).
Research funding at a glance
12 sources graded. The counts below are the independence tiers assigned in the scorecard above. They count sources, not funding amounts or evidence quality.
Public funding alone does not prove independence, and commercial funding alone does not prove a claim false. Named funders, countries and relationships are in the scorecard and source notes.
Related research
- Lion’s Mane: Cognition, Mood, Preparations and Safety Evidence
- Omega-3: Fish Oil, Krill, Algal — Which Form Actually Works
- Omega-3, CoQ10 and Heart Supplements: Evidence by Condition and Funding
- Anticoagulants and Bleeding Safety: Warfarin, DOACs and Monitoring
- Supplements and Heart Medicines: Blood Thinners, Potassium, Herbs and Test Interactions
- Stress, Anxiety and Depression Supplements: Evidence and Safety
- Ashwagandha: Evidence for Stress, Sleep and Safety
- Creatine Monohydrate: Strength, Cognition, Safety and Funding
- The COA Transparency Audit: How 20 Supplement Brands Score on Publishing Lab Results
- The Human-Trial Evidence Index: 25 Supplement Ingredients Ranked by Human-Trial Evidence
Frequently asked questions
Does ginkgo prevent dementia?
No. The two largest trials, GEM (3,069 people) and GuidAge (2,854 people), found no reduction in dementia or Alzheimer's disease (GEM; GuidAge). NCCIH reaches the same conclusion (NCCIH).
Does ginkgo improve memory in healthy people?
Not reliably. NCCIH describes the research as generally low quality, and a systematic review found inconsistent, imprecise and often flawed studies (NCCIH; Crawford 2021).
Is ginkgo safe with warfarin or other blood thinners?
Ask your prescriber or pharmacist. The EMA says warfarin studies do not indicate an interaction, but advises monitoring and says people on anticoagulants or antiplatelets should use ginkgo only after consulting a doctor (EMA monograph).
Should ginkgo be stopped before surgery?
The EMA monograph advises stopping it 3 to 4 days before surgery as a precaution (EMA). A review of supplements and bleeding suggests stopping non-essential supplements 2 weeks before (Hatfield 2022). Tell the surgical team what you take.
Does bacopa work?
It may slightly improve some memory recall or attention measures after about 12 weeks, but the trials are few, short and rated differently for bias by different reviews, and we could not verify who funded the pooled trials (Pase 2012; Kongkeaw 2014; Cave 2023).
What did the FTC say about Prevagen?
In 2017 the FTC and New York State alleged its memory claims were false and unsubstantiated (FTC). In December 2024 the FTC said the court ruled in its favour and ordered the company to stop the challenged claims (FTC statement).
Is citicoline a proven memory supplement?
No. The Cochrane review found some short-term benefit in older people with chronic cerebral disorders, mostly in short trials with varied designs (Cochrane 2005).
Are the trials funded by supplement makers unreliable?
Not automatically. Funding is a reason to look more closely, not proof of a false result. GuidAge, funded by a drug company, still reported a null result (GuidAge).
Sources and funding notes
What we could not verify: funding statements for Pase 2012, Kongkeaw 2014, Crawford 2021 and the 2005 citicoline Cochrane review (journal pages were blocked or lacked a statement); the Cochrane ginkgo review's "sources of support" section; the text of the FDA phosphatidylserine letters; any independent assay of ginkgo or bacopa shop products; the ownership of Schwabe, Ipsen and Quincy; NICE dementia guidance on ginkgo (page blocked); GEM's secondary report on cognitive decline; bacopa liver and interaction data; the court's full reasoning in the Prevagen case; and NTP dose levels. This session's web-search tool was exhausted, so the evidence base is not a full systematic search. Chinese trials, newer ginkgo studies and manufacturer-funded studies of newer extracts may exist that we did not review.
- NCCIH: Ginkgo — US federal fact sheet; US government funded; co-funder of GEM.
- DeKosky et al., JAMA 2008 (PubMed) and full text — GEM randomised trial; NCCIH, NIA and other NIH grants; Schwabe donated product.
- EMA/HMPC EU herbal monograph on Ginkgo biloba leaf (2015), assessment report and product page — EU regulator; public funding.
- Birks & Grimley Evans, Cochrane 2009 (PubMed) and full text — systematic review; Alzheimer's Society supported a previous version; many included trials funded by Schwabe.
- Vellas et al., Lancet Neurology 2012 (GuidAge) — randomised trial; funded by Ipsen.
- Cave et al. 2023 (PubMed) and full text — systematic review; NHMRC grants and a FIT-BioCeuticals PhD scholarship; funders had no role per authors.
- Crawford, Boyd & Deuster 2021 — systematic review of cognitive supplements in healthy adults; funding not verified.
- Pase et al. 2012 — bacopa systematic review; funding not verified.
- Kongkeaw et al. 2014 — bacopa meta-analysis; funding not verified.
- Cochrane review of CDP-choline (2005) — systematic review; funders of included trials not checked.
- Liao et al. 2020 — ginkgo meta-analysis in Alzheimer's disease, not graded in the scorecard; funding not verified.
- Hatfield, Saad & Housewright 2022 — review of supplements and bleeding; no funding statement found in the record.
- US National Toxicology Program, TR-578 abstract — government animal carcinogenicity study.
- FTC press release, 9 January 2017, FTC case page and FTC statement, 10 December 2024 — US federal enforcement; government funded.
- FDA: Unproven Alzheimer's disease products, FDA warning letter to Peak Nootropics LLC and FDA qualified health claims letters — US regulator; government funded.
- Newmaster et al., BMC Medicine 2013 (retracted 2024) — cited only to note its retraction; funded by Canadian research grants per its acknowledgements.
Educational research reviewed 6 October 2026. This article is not medical advice; talk to a qualified clinician or pharmacist before starting, stopping or combining treatments.
Have a question — or want us to cover something?
Ask about anything on this page, or request the next deep dive: an ingredient, a supplement, or a health concern. We use published research, evidence syntheses, and regulatory guidance, with clear source links.
One daily research roundup
Get the topics, key findings and links from our new articles in one email. At most one digest a day, only when there is something new.