# Liver–metabolic stress, apolipoprotein E ε4, and cognition and amyloid burden: findings from the dementia platform Korea trial-ready registry

**Authors:** Young Hyeon Ahn, Jin Gu Kang, Dahyeon Choi, Dae-Jin Kim, Chan-Mo Yang, Sang-Yeol Lee, Sae Hwan Lee, Young Hyun Jung, Sung-Hoon Yoon

PMC · DOI: 10.3389/fnagi.2026.1773977 · Frontiers in Aging Neuroscience · 2026-03-11

## TL;DR

Liver-metabolic stress and APOE ε4 allele dose are linked to cognitive decline and amyloid burden, with interactions varying by liver health stages.

## Contribution

The study reveals how liver-metabolic indices interact with APOE ε4 to influence cognition and amyloid burden in older adults.

## Key findings

- Higher TyG index and AST/ALT ratio are associated with lower MMSE scores and memory performance.
- APOE ε4-related cognitive differences are modified by FIB-4 stage, showing a cross-over pattern.
- Amyloid PET SUVR differences by APOE ε4 are more pronounced at high FIB-4 stages.

## Abstract

Liver–metabolic stress and apolipoprotein E (APOE) ε4 are implicated in late-life cognitive vulnerability, yet how hepatic–metabolic indices relate to cognition and amyloid burden and whether these associations vary by APOE ε4 allele dose remains unclear. We examined liver–metabolic indices in relation to cognition and amyloid PET SUVR and tested effect modification by APOE ε4.

We analyzed baseline data from the Dementia Platform Korea Trial-Ready Registry (DPK-TRR). Primary multivariable analyses used complete cases for outcomes and covariates (n = 507); amyloid PET analyses used the PET subset (n = 496). Exposures included the triglyceride–glucose (TyG) index, AST/ALT ratio, and Fibrosis-4 (FIB-4) stage (low/intermediate/high). APOE ε4 was modeled as allele dose (0/1/2). Multivariable linear regression evaluated associations with global cognition (MMSE), selected domain outcomes, and amyloid PET SUVR, including FIB-4 stage × APOE ε4 dose interaction terms, adjusted for age, sex, education, hypertension, diabetes, and dyslipidemia.

Higher TyG index and AST/ALT ratio were associated with lower MMSE scores (TyG: β = −1.13, 95% CI − 2.11 to −0.16; p = 0.022; AST/ALT: β = −1.40, 95% CI − 2.10 to −0.39; p = 0.007) and with memory performance on CERAD delayed recall (TyG: β = −0.29, 95% CI − 0.52 to −0.06; p = 0.012), whereas neither marker showed a clear association with amyloid PET SUVR. The association between FIB-4 stage and cognition differed by APOE ε4, showing a cross-over pattern across fibrosis stages: APOE ε4–associated differences were attenuated and reversed at intermediate/high FIB-4 compared with low FIB-4. In the amyloid PET subset, APOE ε4 group differences in SUVR were not prominent at low FIB-4, but tended to diverge at high FIB-4 with higher SUVR in groups with greater ε4 burden.

Routine liver–metabolic indices were associated with cognitive performance, while FIB-4 stage showed effect modification by APOE ε4 in relation to both cognition and amyloid PET SUVR. These findings support heterogeneity in liver–metabolic and genetic contributions to late-life cognitive vulnerability in a dementia trial-ready registry and motivate longitudinal studies to clarify temporal relationships.

## Linked entities

- **Genes:** APOE (apolipoprotein E) [NCBI Gene 348]

## Full-text entities

- **Genes:** APOE (apolipoprotein E) [NCBI Gene 348] {aka AD2, APO-E, ApoE4, LDLCQ5, LPG}, SLC17A5 (solute carrier family 17 member 5) [NCBI Gene 26503] {aka AST, ISSD, NSD, SD, SIALIN, SIASD}
- **Diseases:** dyslipidemia (MESH:D050171), Dementia (MESH:D003704), Fibrosis (MESH:D005355), hypertension (MESH:D006973), amyloid (MESH:C000718787), diabetes (MESH:D003920)
- **Chemicals:** TyG (-), triglyceride (MESH:D014280), glucose (MESH:D005947)

## Full text

_Full body text omitted from this summary view._ Fetch the complete paper as Markdown: https://tomesphere.com/paper/PMC13012996/full.md

## Figures

4 figures with captions in the complete paper: https://tomesphere.com/paper/PMC13012996/full.md

## References

63 references — full list in the complete paper: https://tomesphere.com/paper/PMC13012996/full.md

---
Source: https://tomesphere.com/paper/PMC13012996