← The Signal Report Work with me

Report 254 · Supplement Reality Check

Does ginkgo prevent memory loss?

Ginkgo biloba is one of the oldest "brain health" supplements on the shelf, and it is unusual in one way: it has been tested for dementia prevention in two large, long, placebo-controlled trials. Nearly 6,000 older adults took a standardized extract or a placebo for five to six years. Neither trial found that ginkgo prevented dementia. Here is what they measured, what they didn't, and where ginkgo's evidence still has some life in it.

Disclosure up front: I run a functional-mushroom supplement company, so cognitive-health supplements are a market I have a commercial stake in. We don't sell ginkgo, and I have no connection to any study discussed here. The full disclosure is at the bottom.

The big one: GEM

The Ginkgo Evaluation of Memory (GEM) study was funded by the U.S. National Institutes of Health and published in JAMA in November 2008. Steven DeKosky and colleagues randomized 3,069 community volunteers aged 75 or older at five U.S. academic medical centers: 2,587 had normal cognition and 482 had mild cognitive impairment (MCI). Half took 120 mg of a standardized ginkgo leaf extract (EGb 761) twice a day; half took an identical placebo. They were checked every six months for a median of 6.1 years.

The dose and product were not a weak test. The paper says the 240 mg daily dose was chosen because earlier studies suggested a dose-response up to that level, EGb 761 is used in many branded U.S. ginkgo products, and an independent lab confirmed the contents of each batch of active and placebo tablets. The manufacturer, Schwabe, donated the tablets and placebo; the paper says it had no role in the design, analysis or writing.

What GEM found

523 people developed dementia: 277 on ginkgo and 246 on placebo. That works out to 3.3 versus 2.9 cases per 100 person-years. The hazard ratio for all-cause dementia was 1.12 (95% confidence interval 0.94 to 1.33), and for Alzheimer's disease 1.16 (0.97 to 1.39). In people who started with MCI, ginkgo did not slow progression to dementia either (hazard ratio 1.13, 0.85 to 1.50). None of those differences was statistically significant, and the point estimates lean slightly the wrong way. The authors' conclusion was blunt: ginkgo "cannot be recommended for the purpose of preventing dementia."

A second GEM paper (Snitz and colleagues, JAMA, December 2009) asked the softer question many buyers actually care about: does it slow everyday cognitive decline, even short of dementia? Across memory, attention, visual-spatial skills, language and executive function, the yearly rates of decline did not differ between ginkgo and placebo, and no subgroup (by age, sex, education, APOE4 status or baseline MCI) stood out.

The second trial agreed

GuidAge, a French trial published in The Lancet Neurology in 2012, came at the question from a different angle: adults aged 70 or older who had gone to their family doctor with memory complaints, which is closer to the person who actually buys ginkgo. Vellas and colleagues randomized 2,854 people to the same 120 mg twice-daily EGb 761 or placebo for five years. Probable Alzheimer's was diagnosed in 61 people on ginkgo and 73 on placebo (1.2 versus 1.4 per 100 person-years; hazard ratio 0.84, 0.60 to 1.18, p = 0.306). The authors concluded ginkgo did not reduce the risk of progressing to Alzheimer's. The abstract adds that the risk "was not proportional over time," a statistical caveat about how the two groups' curves behaved. GuidAge was funded by Ipsen.

The U.S. National Center for Complementary and Integrative Health, which co-funded GEM, now summarizes it plainly: ginkgo "has not been shown to be beneficial for preventing or slowing the progression of dementia."

What these trials can't rule out

Timing. GEM's participants were 79 on average when they started. The authors note that the brain changes behind dementia begin years before symptoms, so a supplement started earlier in life might behave differently. That's a fair caveat, but it is a hypothesis nobody has tested, not evidence that earlier use works.

Adherence. By the end of GEM, 60.3% of the remaining participants were still taking their study tablets, equally in both groups. Real-world adherence is rarely better, so this is a realistic test of "take it for years," but it does dilute any effect a perfectly adherent person might get.

A subgroup to treat with caution. In an exploratory analysis, GEM participants with a history of cardiovascular disease who took ginkgo had a higher dementia rate (hazard ratio 1.56). The authors call this "puzzling," say it "should be viewed with caution," and note there's no known mechanism. I'd file it the same way: an unplanned subgroup finding, not evidence of harm.

Safety, briefly

GEM found similar adverse-event and death rates in both groups, and no difference in major bleeding (hazard ratio 0.97). Hemorrhagic strokes were twice as common on ginkgo, 16 versus 8, but the numbers were small and not statistically significant; the authors say it "should be explored in future studies." GEM excluded people on warfarin, and NCCIH notes ginkgo may raise bleeding risk in people taking anticoagulants and may interact with other drugs.

Where ginkgo still has data: existing dementia

Prevention in healthy people and treatment of people who already have symptoms are different questions, and marketing tends to blur them. A Cochrane review published in February 2026 (Wieland and colleagues, 82 studies, 10,613 participants) looked only at people with cognitive complaints, MCI or dementia. Its conclusions, from the abstract:

For MCI, moderate-certainty evidence says ginkgo "probably has little or no effect" at six months on overall clinical status, cognition or daily activities. For diagnosed dementia, low-certainty evidence suggests there "may be small to moderate benefits" at six months. Those dementia pools are very inconsistent: the cognition estimate (a 1.86-point improvement on a short cognitive test, 95% CI 0.24 to 3.48) had heterogeneity (I²) of 96%, which means the individual trials disagree a lot. Only four of the 82 studies were at low risk of bias in every domain. That is a maybe for people already diagnosed, under medical care, not a reason for a healthy 50-year-old to take it for memory.

How I'd read it

Ginkgo is one of the few supplements where the big, expensive, long-term prevention trial was actually run, with a standardized extract at a meaningful dose. Twice. The honest summary: in nearly 6,000 older adults followed for five to six years, ginkgo did not prevent dementia or slow memory decline. The small, uncertain signal that remains is in people who already have dementia, and that is a conversation for their doctor.

I'd apply the same test to any "supports memory" label, including the ones in my own industry: has it been tested on the outcome you care about, in people like you, for long enough? It's the same lens I used for DHA and memory and the one laid out in what "clinically proven" really means.

What would change my read: a well-run trial showing a prevention benefit in a younger group, or a high-certainty update to the Cochrane dementia findings. If one lands, I'll update this post in place.

Not medical advice. This is educational analysis, not a recommendation — a study is not a prescription. Talk to a qualified clinician before acting on anything you read here. Full disclaimer →

Sources

  1. Steven T. DeKosky, Jeff D. Williamson, Annette L. Fitzpatrick, et al., for the GEM Study Investigators, "Ginkgo biloba for Prevention of Dementia: A Randomized Controlled Trial," JAMA 300(19):2253–2262 (19 November 2008), DOI 10.1001/jama.2008.683; author manuscript on PMC2823569; ClinicalTrials.gov NCT00010803. (Primary source. Full author manuscript read via PubMed Central; online eTables not read. Source of the design, dose rationale, batch testing, Schwabe's role, participant numbers, dementia counts and rates, hazard ratios, adherence, the cardiovascular subgroup, the safety table and the quoted phrases.)
  2. B. E. Snitz, E. S. O'Meara, M. C. Carlson, et al., "Ginkgo biloba for Preventing Cognitive Decline in Older Adults: A Randomized Trial," JAMA 302(24):2663–2670 (23 December 2009), DOI 10.1001/jama.2009.1913; PMC2832285. (Abstract read via Europe PMC. Source of the no-difference finding on rates of decline by domain and the absence of subgroup effects.)
  3. Vellas, Coley, Ousset, Berrut, et al., "Long-term use of standardised ginkgo biloba extract for the prevention of Alzheimer's disease (GuidAge): a randomised placebo-controlled trial," The Lancet Neurology 11:851–859 (October 2012), DOI 10.1016/S1474-4422(12)70206-5; ClinicalTrials.gov NCT00276510. (Abstract read via Europe PMC; metadata via Crossref. Source of the population, numbers, hazard ratio, the non-proportional-risk caveat and the Ipsen funding.)
  4. L. S. Wieland, E. Ludeman, Y. Chi, et al., "Ginkgo biloba for cognitive impairment and dementia," Cochrane Database of Systematic Reviews 2026, Issue 2, CD013661 (5 February 2026), DOI 10.1002/14651858.CD013661.pub2. (Abstract read via Europe PMC; full review not read. Found through a Consensus literature search. Source of the 82 studies and 10,613 participants, the MCI and dementia conclusions, the cognition estimate and I², the four low-risk-of-bias studies and the quoted phrases.)
  5. National Center for Complementary and Integrative Health, "Ginkgo" fact sheet (last updated February 2025). (Read. Source of the quoted summary on dementia and the anticoagulant and drug-interaction cautions.)

"Nearly 6,000" adds GEM's 3,069 to GuidAge's 2,854 randomized participants (5,923, my arithmetic). The Consensus search also surfaced two meta-analyses of EGb 761 in dementia (Tan et al. 2014; Riepe et al. 2025); they were read in abstract only, cover treatment rather than prevention, and are not cited above.

Disclosure, plainly: I founded and run Shroombiosis (a company I run), which formulates and sells functional-mushroom supplements, so I have a commercial stake in how people think about supplements and brain health. We don't sell ginkgo, and I have no relationship with any study, author or manufacturer discussed here. Nothing here is sponsored and no link earns a commission; here's the full policy. A recommendation with no stake at all: for performance nutrition, Die Tryin Co. is a fellow combat-veteran-owned brand I'm glad to point people to. I don't own it and earn nothing from the link.

Onur Oncer
Onur Oncer

U.S. Army combat veteran (Counter-IED / Electronic Warfare), peer-reviewed researcher in microwave spectroscopy, and founder & CEO of Shroombiosis. Consults on laboratory operations, AI, and supplement formulation.

← All reports