31/08/2026
A new generation of Alzheimer biomarkers is changing how we investigate dementia prevention, and this has delivered significant implications for reinterpreting the role of Ginkgo biloba in this context.
A 2026 Korean study followed 35 people with amyloid PET-positive mild cognitive impairment (MCI) for 18 months: 21 received standardised Ginkgo extract 240 mg/day and 14 received other cognitive enhancers. All Ginkgo-treated patients remained clinically stable, compared with cognitive decline in 57.1% of the comparator group. None receiving Ginkgo progressed to Alzheimer dementia, versus 28.6% of comparators.
More significantly, plasma amyloid-β oligomerisation decreased and amyloid PET burden remained stable with Ginkgo, whereas both measures worsened without it. Although small and retrospective, this is groundbreaking research because it links a herbal intervention to longitudinal changes in both a blood biomarker and serial brain imaging, rather than relying solely on cognitive tests.
This work advances a new model for investigating whether herbs can slow the biology that eventually might produce dementia. Biomarkers such as amyloid and tau PET, plasma phosphorylated tau-217, neurofilament light chain, glial fibrillary acidic protein and MRI-measured brain atrophy can reveal biological change well before dementia becomes apparent. Ginkgo is the first herb to be studied in this way, but the approach could readily be applied to other promising herbs and combinations.
This study actually represents the latest stage of a coherent research program from the same Korean group. In 2025, they reported that adding Ginkgo to donepezil in 101 amyloid PET-confirmed Alzheimer patients improved Mini-Mental State Examination scores by 2.4 points and reduced plasma amyloid-β oligomerisation more than donepezil alone. Specifically, MDS-OAβ (Multimer Detection System-Oligomerised Amyloid-β) measures how added amyloid-β forms oligomers in a patientās plasma, providing an indirect blood-based index of amyloid-aggregation tendency that has been associated with cerebral amyloid burden, but does not directly measure brain oligomers.
A second earlier study in 64 amyloid-positive MCI patients found that Ginkgo preserved cognition and daily function, reduced oligomerisation and was associated with no conversions to dementia, compared with 13.6% among patients receiving other cognitive enhancers. A subsequent pooled analysis of 165 amyloid-positive patients again associated Ginkgo exposure with better cognitive and functional trajectories and lower amyloid oligomerisation. The consistency is striking, although all studies were retrospective, originated from the same research network and await independent replication.
This provides an important new context for evaluation of the earlier two large negative Ginkgo prevention trials. The Ginkgo Evaluation of Memory (GEM) study followed 3,069 adults aged 75 years or older for a median of 6.1 years but found no reduction in dementia or Alzheimer disease at 240 mg/day of extract. Similarly, the GuidAge study followed for five years 2,854 adults aged over 70 with memory complaints without finding a significant reduction in progression to Alzheimer disease. These negative findings remain valid, but both trials depended primarily on the eventual appearance of clinical dementia, enrolled biologically heterogeneous populations and began treatment quite late in life. Given that Alzheimer pathology develops over decades, even five or six years may be insufficient in older patients for a modest slowing of pathology to translate into fewer dementia diagnoses.
Modern biomarkers allow us to ask the question earlier and more directly: not whether fewer people have crossed over clinical threshold for dementia, but instead whether a herb is actually slowing the underlying disease process. This appears to be what Ginkgo is doing based on the Korean research.
These papers represent an interesting evolution in the Ginkgo story. Rather than studying unselected patients with āpossible Alzheimer disease risk,ā they focus on patients with amyloid PET-confirmed prodromal Alzheimer disease, thereby reducing one of the major weaknesses of previous Ginkgo prevention trials, namely diagnostic heterogeneity.
Although MDS-OAβ remains less widely accepted than the biomarkers within the AT(N) framework, this limitation is partly offset by the use of amyloid PET both to confirm underlying Alzheimer pathology and, in the 18-month study, to measure changes in cerebral amyloid burden over time. The findings therefore do not rest solely on this exploratory plasma marker, although replication using more established measures such as plasma phosphorylated tau-217 and neurofilament light chain would strengthen them.
The AT(N) framework: A = amyloid biomarkers (PET or CSF/plasma Aβ42/40), T = phosphorylated tau (particularly plasma or CSF p-tau217) and N = neurodegeneration (MRI, fluorodeoxyglucose positron-emission tomography (FDG-PET) or neurofilament light chain (NfL)) is being used in most major pharmaceutical trials.
Another key weakness of the Korean studies is their retrospective, non-randomised design, which cannot exclude treatment-selection bias or unmeasured confounding. However, the clinical signal is unusually strong and consistent across cognition, daily functioning and biomarkers.
For more information see:
https://pubmed.ncbi.nlm.nih.gov/41984497/
https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2025.1563056/full
https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2025.1639924/full
https://link.springer.com/article/10.1186/s12906-026-05469-1