Ginkgo Biloba: Effects & Evidence — what is supported

Ginkgo biloba (often as the standardized extract EGb 761) is one of the best-known herbal preparations in the areas of cognition and circulation/vascular health. However, the evidence base is strongly indication-dependent: meta-analyses show relatively consistent effects for some conditions, while for others (e.g., Alzheimer’s) the overall picture is more heterogeneous and sometimes limited by predominantly preclinical data.

Below, I’ll separate what is supported by evidence, where the data are limited, and what you should consider in practice—including safety and interaction caveats.

First lifestyle: what to check before Ginkgo biloba

If your goal is cognition, mood, or inflammation, the most promising levers are usually sleep, movement, light exposure during the day, and overall nutrition quality—often measurable faster than relying on a supplement alone. For hearing problems, there’s an additional point: time matters, and it does not replace medical evaluation.

So start with the question: Why do you want to take Ginkgo in the first place? Example (cognition): If sleep quality is poor, memory and attention are often the most strongly affected. In practice, better sleep hygiene (consistent times, less late light, reducing alcohol late in the day) often yields measurable improvements before a supplement is even considered. For many “inflammation” or “stress sensitivity” goals, the same idea applies: light exposure during the day and regular activity influence both fatigue and stress physiology, which can indirectly shift results on “cognitive” or “biological” endpoints.

For hearing problems, the most important point is not “which product,” but urgency: sudden sensorineural hearing loss is a time-critical event. In meta-analyses, Ginkgo is discussed as add-on therapy (Yuan et al., 2023, PMID 36999358; Si et al., 2022, PMID 34474140), but that does not replace acute care. If you suspect something like this (e.g., clearly worse hearing in one ear over a short period, a feeling of pressure, or associated symptoms), you need early medical diagnostics—because the therapeutic time window can be clinically relevant.

Finally: bleeding risk and cardiovascular medications. In the overviews mentioned, Ginkgo is mostly evaluated as an extract under study conditions; in real life, additional risk factors and interactions may matter. If you take anticoagulants or have a relevant cardiovascular condition, you shouldn’t “just test” Ginkgo—discuss it with a clinician first.

If you want a more systematic lifestyle-first approach, this context may also help: Stress resilience: Effects & evidence — what is actually supported.

What quality of Ginkgo matters: extracts are not all the same

The products evaluated in the relevant meta-analyses are usually standardized Ginkgo biloba extracts, especially EGb 761. This is crucial: “Ginkgo” is not the same as “Ginkgo,” and the observed effects therefore apply more to specific extracts and investigated patient groups—not automatically to every commercial product.

Why does this matter? Herbal extracts vary by raw material origin, extraction process, standardization, and the content ranges of characteristic constituents. That can change bioavailability, concentration ratios, and pharmacological activity. If a meta-analysis explicitly evaluates EGb 761 or “standardized Ginkgo biloba extracts,” that does not mean that an arbitrary Ginkgo product will produce the same effect.

This is reflected in how the evidence must be interpreted: For the indication mild dementia, Riepe et al. 2025 concludes that the EGb 761 extract may be safe and effective (Riepe et al., 2025, PMID 39895346). That’s a concrete statement about the studied extract class and about the included RCT data. Extrapolating to “any Ginkgo” would be methodologically too far.

The same extract-specificity is relevant for other indications. For sudden sensorineural hearing loss, Ginkgo is addressed in meta-analyses as add-on therapy (Yuan et al., 2023, PMID 36999358; Si et al., 2022, PMID 34474140). The same applies for unstable angina pectoris, where Liu et al. 2024 evaluated efficacy and safety in a systematic review and network meta-analysis (Liu et al., 2024, PMID 38718890). Even for blood metabolomics in type-2 diabetes, Zou et al. 2023 refers to the extract interventions used in the studies (Zou et al., 2023, PMID 38264278).

Practical decision-rule: when selecting a product, specifically look for standardization details and check whether they match the extracts used in studies (e.g., “EGb 761”). The evidence base is therefore not just “Ginkgo yes/no,” but a question of fit.

If you want a general framework for the logic of how it might work, Immune modulation: Effects & evidence — what is supported can also help, because it similarly separates plausible mechanisms from robust clinical endpoints.

Evidence hierarchy: why meta-analyses carry more weight (and what remains uncertain)

Meta-analyses pool data from multiple studies, so they are often more reliable than individual trials. The key advantage is that they increase sample size and can make inconsistent findings more visible. Still, limitations remain: differences in study design, patient populations, extracts, and endpoints can restrict how strong the conclusions can be.

Meta-analyses are stronger because they synthesize results systematically. While a single RCT may yield few events (e.g., for rare side effects or heterogeneous patient groups), the pooled analysis can reveal patterns. That is exactly why the indication-specific claims in this article are primarily tied to meta-analyses:

• Mild dementia: Riepe et al. 2025 (PMID 39895346) evaluates EGb-761 data across patient subgroups in randomized studies.
• Sudden hearing loss: Yuan et al. 2023 (PMID 36999358) and Si et al. 2022 (PMID 34474140) assess meta-evidence for add-on therapy.
• Unstable angina pectoris: Liu et al. 2024 (PMID 38718890) combines efficacy and safety data in a systematic review plus network meta-analysis.
• Type-2 diabetes/blood metabolism: Zou et al. 2023 (PMID 38264278) evaluates effects and safety on blood-metabolism-related endpoints.

But: meta-analysis is not “universal truth.” First, the included studies differ in quality, dosing, and duration (you need to check the specifics in each paper). Second, generalizability is limited if the population studied doesn’t match yours (age, disease severity, comorbidities). Third: “safety” in overviews is often described as no meaningful increase in certain events or as “tolerable within the studied context”—that is not the same as absolute risk-free use for an individual.

For mechanistic clues, there’s an additional caveat: lab or animal data can make it plausible why a compound might work (e.g., neuroprotective or antioxidative mechanisms). But they do not replace evidence of efficacy in humans. The same idea is emphasized for other indications—such as spinal cord injury, where Wu et al. 2025 provides a meta-analysis and systematic overview of experimental animal data and possible mechanisms (Wu et al., 2025, PMID 40019748). That supports understanding, but does not directly answer the question “does it reliably work in humans?”

If you want to go deeper into methodology: the most important practical rule is to interpret meta-analyses by indication and extract. That reduces the risk of the false inference “an effect somewhere” → “it works everywhere.”

Effects with meta-analysis evidence: where there are hints of benefit

The evidence base for Ginkgo biloba (especially standardized extracts like EGb 761) provides indications where there are indeed hints of benefit. The strongest pattern is where meta-analyses analyze RCT data or systematically pooled clinical endpoints—for example, in mild dementia, as add-on treatment for sudden sensorineural hearing loss, and in the context of unstable angina pectoris. For other goals, the picture is much less consistent.

Mild dementia

Riepe et al. 2025 (PMID 39895346) concludes in a meta-analysis that the EGb 761 extract may be safe and effective in mild (i.e., mild-grade) dementia (Riepe et al., 2025, PMID 39895346). Important: this is not “for every type of memory problem,” but tied to a specific indication. Also key: this is a meta-analysis of RCT data—this evidence class is stronger than arguments based purely on mechanistic reasoning.

Sudden sensorineural hearing loss (as add-on therapy)

For sudden sensorineural hearing loss, meta-analyses evaluate Ginkgo as add-on therapy. Yuan et al. 2023 (PMID 36999358) and Si et al. 2022 (PMID 34474140) discuss efficacy and safety within the included study framework (Yuan et al., 2023, PMID 36999358; Si et al., 2022, PMID 34474140). The methodological key point: “add-on therapy” means Ginkgo is not tested in isolation but alongside a base treatment. Therefore, you shouldn’t interpret these results as a substitute for standard care.

Unstable angina pectoris

Liu et al. 2024 (PMID 38718890) reports in a systematic review and network meta-analysis on efficacy and safety of a Ginkgo extract for unstable angina pectoris (Liu et al., 2024, PMID 38718890). Network meta-analyses are particularly useful when there are multiple comparison arms or different therapy variants. Again, generalizability remains tied to specific study settings.

Blood metabolism in type-2 diabetes

Zou et al. 2023 (PMID 38264278) evaluates efficacy and safety of Ginkgo on blood-metabolism-related endpoints in type-2 diabetes (Zou et al., 2023, PMID 38264278). This is relevant because “diabetes” as an umbrella term includes many endpoints (e.g., glucose parameters, lipids, etc.). Meta-analyses can help sort inconsistent single-study results here—but they don’t guarantee clinical outcomes like “fewer complications” when trials mostly measure surrogate endpoints.

Note on dosing: In this article, I do not provide a blanket dosing value because the exact dosing and timing can differ across indications in the respective meta-analyses. Providing specific numbers without the detailed dosing information from the underlying studies in the referenced meta-analyses would not be adequately supported. If you want, I can take the next step and extract dosing ranges per indication from the relevant papers—but that would require access to the full texts/details of those meta-analyses.

Limitations & “still unclear”: Alzheimer and spinal cord injury do not directly mean “it works”

For some goals, the evidence strength for Ginkgo is substantially less stable. For Alzheimer’s, the overall evidence is heterogeneous; for spinal cord injury, the meta-level evidence is mostly based on animal data and mechanisms, leaving open the step toward clinical effectiveness. For inflammatory markers, there may be effects on lab values—yet that doesn’t automatically translate into improved clinical symptoms or endpoints.

Alzheimer’s

Xie et al. 2022 (PMID 35159288) discusses whether Ginkgo extract can be used in Alzheimer’s—with “lessons” drawn from preclinical and clinical studies (Xie et al., 2022, PMID 35159288). The key takeaway must be read carefully: even if there are theoretical/mechanistic reasons and some studies show effects, the overall conclusion is not so clear that you could derive a robust, indication-stable recommendation for “Alzheimer’s in general.” This limitation matters: Alzheimer’s is not only “memory”—it’s a complex disease with multiple endpoints. When the meta-analysis synthesizes the evidence across different study levels, it often remains unclear how consistent and clinically meaningful the effects truly are.

Spinal cord injury

Wu et al. 2025 (PMID 40019748) provides a meta-analysis and systematic overview of neuroprotective effects and possible mechanisms in experimental animal models (Wu et al., 2025, PMID 40019748). This strengthens mechanistic understanding, but: animal data are not a clinical guarantee in humans. So if you want to use Ginkgo for spinal cord-related issues, the honest evidence level is closer to “plausible/experimental” than “clinically established” (in humans).

Inflammatory markers

Mousavi et al. 2022 (PMID 35781715) evaluates in a systematic review and meta-analysis the effects on inflammation-related markers in clinical studies (Mousavi et al., 2022, PMID 35781715). The key methodological point: inflammatory markers are biomarkers. Biomarkers may improve without automatically leading to noticeable clinical improvement (or vice versa). Also, practical relevance depends on baseline levels, disease activity, and whether studies focus on surrogate vs. hard endpoints.

In short: not every promising lab or preclinical finding becomes clinically effective. This is particularly true for complex neurodegenerative diseases and for targets operationalized mainly as biomarkers in studies.

If you set up your decision logic “disease-specific,” the clear indication boundaries from stronger meta-analyses help—and you can then align them with lifestyle levers that often have broader impact than any single supplement.

Table: overview of the existing evidence by indication

What to take away from it (Bottom Line)

• Indication before product: The best evidence is where meta-analyses include RCT data or systematic clinical syntheses (e.g., mild dementia with EGb 761: Riepe et al., 2025, PMID 39895346).
• No one-size-fits-all: “Ginkgo works” is too general—the effects are extract- and target-specific (e.g., add-on therapy for hearing loss: Yuan et al., 2023, PMID 36999358; Si et al., 2022, PMID 34474140).
• Lifestyle first, especially for cognition/mood/inflammation: Sleep, movement, light, and nutrition are often faster and broader levers than a single supplement.
• Don’t delay acute symptoms: For sudden hearing loss, early medical evaluation is crucial—Ginkgo can only be discussed as add-on in meta-analyses, not as a replacement.
• State evidence gaps honestly: For goals like Alzheimer’s, the data are overall more heterogeneous; for spinal cord injury, the evidence is sometimes strongly based on animal/mechanistic data (Xie et al., 2022, PMID 35159288; Wu et al., 2025, PMID 40019748).

If you want, as the next step I can create an indication-based decision checklist (including typical study durations/endpoints per indication). To do that, tell me which goal is most central for you.