Neurology · Medical Literature
A reappraisal of APOE genetic effects on Alzheimer's disease risk in the Japanese population: a meta-analysis.
Abstract
Our meta-analysis demonstrates that the risk effect of APOE-e4*4 relative to APOE-e3*3 for Alzheimer's disease in the Japanese population is approximately 12-15-fold, comparable to that reported in Caucasian populations, rather than greater than 20-fold as previously reported.
Clinical relevance
Among Japanese individuals, is the APOE-ε4/ε4 (homozygous) genotype associated with a greater magnitude of Alzheimer's disease risk compared with APOE-ε3/ε3, and does this risk differ meaningfully from that observed in Caucasian populations? __SNAPSHOT__ {"studyDesign":"Systematic review with meta-analysis of case-control studies","setting":"Japan; multiple institutions; literature searches conducted via PubMed and Web of Science (WOS) Core Collection","sampleSize":"21 eligible case-control studies included in the primary meta-analysis (early-onset AD: n=8 studies; late-onset AD: n=11 studies; AD without age stratification: n=10 studies); individual patient counts not reported in the abstract","population":"Japanese individuals from case-control studies reporting APOE genotype data in both Alzheimer's disease cases and controls; spanning studies from the early 1990s to the 2010s; included both early-onset and late-onset AD populations","intervention":"APOE-ε4/ε4 homozygous genotype (exposure)","comparator":"APOE-ε3/ε3 homozygous genotype (reference group)","primaryOutcome":"Odds ratio (OR) for Alzheimer's disease risk associated with APOE-ε4/ε4 vs. APOE-ε3/ε3 in the Japanese population","followUp":"Not applicable (meta-analysis of cross-sectional case-control studies)","fundingSource":"Not reported in the abstract; senior investigator affiliated with Niigata University"}
Key findings
The pooled meta-analytic OR for APOE-ε4/ε4 vs. APOE-ε3/ε3 was approximately 15.5 for early-onset AD, 12.5 for late-onset AD, and 13.5 for all AD combined, indicating an overall approximately 12–15-fold increased Alzheimer's disease risk for APOE-ε4 homozygotes in the Japanese population. Confidence intervals and p-values for pooled estimates were not reported in the abstract. This contrasts sharply with the landmark 1997 APOE and Alzheimer Disease Meta-Analysis Consortium report, which estimated an OR of 33.1 (95% CI: 13.6–80.5) for Japanese APOE-ε4/ε4 vs. ε3/ε3, and with a subsequent prior meta-analysis range of 21.8–33.1-fold risk increase. The revised estimate of 12–15-fold is now numerically comparable to the risk magnitude reported for Caucasian populations, where the ε4/ε4 OR has been reported at approximately 14.9 (95% CI: 10.8–20.6) relative to ε3/ε3. Adverse events are not applicable to this genetic association meta-analysis. NNT/NNH not calculable from available data.
Practice implications
This meta-analysis meaningfully revises downward the historically reported APOE-ε4/ε4 risk magnitude for Alzheimer's disease in Japanese individuals. The prior dogma—that Japanese APOE-ε4 homozygotes face a disproportionately greater risk (>20- to 33-fold) compared to Caucasian populations—appears to have been an artifact of earlier, smaller, or potentially overlapping cohorts. The authors addressed a critical methodological issue: overlap of subjects across multiple studies using the same Japanese biobank cohorts (JGSCAD and NCGG), which likely inflated prior risk estimates. By including only non-overlapping, large-scale, representative studies, the revised pooled OR converges toward the Caucasian range (~12–15-fold). These findings apply to Japanese adults with late-onset or early-onset AD in case-control designs; population-based prospective incidence data may yield different absolute risk estimates. Clinicians should be cautious about counseling Japanese patients with APOE-ε4/ε4 status using historically cited >20-fold risk figures. IMPORTANT CAVEAT: Only the abstract was available for review; full confidence intervals, heterogeneity statistics (I²), publication bias assessments, funnel plots, and subgroup analyses cannot be evaluated, which limits certainty of interpretation. __LIMITATIONS__ Only the abstract was available; key statistical outputs (individual study ORs, pooled CIs, p-values, I² heterogeneity, Egger/funnel plot analyses) cannot be independently verified.,Inclusion of only English-language studies (14 non-English articles excluded) may introduce language bias, potentially missing relevant Japanese-language data.,All included studies are case-control in design; ascertainment and selection bias inherent to this design cannot be excluded, and absolute population-level incidence risks cannot be derived.,Overlap management between cohorts (JGSCAD, NCGG) required subjective decisions about study inclusion, which itself introduces potential bias in either direction.,Generalizability is limited to the Japanese population; results should not be extrapolated to broader East Asian populations (e.g., Korean, Chinese) without independent data.,Heterogeneity across included studies (differing diagnostic criteria across decades, EOAD vs. LOAD definitions, community vs. clinic-based ascertainment) was likely substantial but not quantifiable from the abstract alone.,Funding sources and conflicts of interest are not reported in the abstract, precluding assessment of sponsorship bias.,The revised estimate aligns APOE-ε4/ε4 risk in Japanese populations with Caucasian populations, but this does not mean absolute risk is equivalent, as baseline APOE-ε4 allele frequency differs across ethnic groups. __BOTTOM_LINE_CATEGORY__ potentially_practice_informing __BOTTOM_LINE__ This meta-analysis of 21 Japanese case-control studies substantially revises prior estimates of APOE-ε4/ε4 Alzheimer's disease risk in Japanese individuals downward—from historically cited >20–33-fold to approximately 12–15-fold—bringing the Japanese estimate in line with Caucasian population data and attributing the prior inflation largely to overlapping study cohorts. Clinicians should update their risk communication for Japanese patients with APOE-ε4 homozygosity accordingly, avoiding citation of the outdated >20-fold figures. Full peer-reviewed confirmation of the statistical methodology (heterogeneity, publication bias) is warranted before definitive practice revision.