Exceptionally low likelihood of Alzheimer’s dementia in APOE2 homozygotes from a 5,000-person neuropathological study

The study investigates whether homozygosity for the APOE2 allele (APOE2/2) confers an especially low risk of Alzheimer’s dementia compared with the already lower-risk APOE2/3 genotype and markedly lower than the high-risk APOE4 genotypes. The context is extensive prior evidence that APOE4 increases risk and lowers age at onset while APOE2 decreases risk and raises age at onset. The authors hypothesized that APOE2/2 would show the lowest odds of Alzheimer’s dementia and that APOE2 and APOE4 gene-dose effects would be stronger in neuropathologically confirmed case-control cohorts than in clinically diagnosed but unconfirmed samples due to diagnostic misclassification.

Prior case-control studies lacking neuropathological or biomarker confirmation may underestimate APOE effects because APOE status influences rates of clinical misclassification; APOE4 non-carriers are more often misclassified. Previous clinically and neuropathologically characterized studies were often underpowered to assess the very rare APOE2/2 genotype (<1% prevalence). The authors’ recent observations of individuals with rare APOE variants (e.g., APOE2 Christchurch) and resistance to clinical onset despite heavy amyloid burden motivated testing whether APOE2 homozygosity markedly reduces late-onset Alzheimer’s dementia risk. The study builds on established associations of APOE with amyloid and tau pathology and seeks to quantify genotype-specific risks using large, well-characterized cohorts.

- Design: Case-control analysis using clinically characterized and neuropathologically confirmed Alzheimer’s dementia cases and cognitively unimpaired controls, compared with a larger neuropathologically unconfirmed clinical cohort; an additional combined dataset was analyzed. - Cohorts: Data assembled by the Alzheimer’s Disease Genetics Consortium (ADGC) from NIA-sponsored centers and studies (e.g., ACT, ROSMAP, ANDI), including over 5,000 neuropathologically confirmed participants and more than 24,000 unconfirmed participants. - Genotyping: APOE genotypes determined at centralized repositories. - Neuropathology: CERAD plaque score (0–3) and Braak neurofibrillary tangle stage used to define and quantify AD pathology; additional pathologies assessed included congophilic amyloid angiopathy (CAA), Lewy body disease (LBD), vascular brain injury (VBI), and hippocampal sclerosis (HS) when available; TDP-43 and microinfarcts were often unavailable and excluded. - Clinical definitions: Alzheimer’s dementia cases met NINCDS/ADRDA criteria; controls lacked dementia and significant AD neuropathology. Age at dementia onset was used when available; age at death was used as proxy when onset age was missing. - Statistical analysis: Logistic regression models estimated odds ratios (ORs), 95% confidence intervals (CIs), and p-values for each APOE genotype relative to reference genotypes (primarily APOE3/3; additional comparisons used APOE2/3 and APOE4/4). Models adjusted for age and sex; allelic dose effects for APOE2 and APOE4 were assessed using additive models. Kaplan–Meier analyses depicted freedom from Alzheimer’s dementia by genotype. Linear models assessed residual effects of APOE genotypes on Braak stage after adjusting for CERAD plaque severity.

- In the neuropathologically confirmed group (primary analysis), odds ratios versus APOE3/3 were: APOE2/2 OR 0.13 (95% CI 0.05–0.36; p=6.3×10^-5), APOE2/3 OR 0.39 (0.03–0.50; p=1.6×10^-12), APOE2/4 OR 2.68 (1.65–4.36; p=7.5×10^-5), APOE3/3 OR 1.63 (0.58–4.71; p≈2.2×10^-13 as reported), APOE4/4 OR 31.22 (16.59–58.75; p=4.9×10^-26). - Allelic dose effects (neuropathologically confirmed): APOE2 dose OR 0.38 (0.30–0.48; p=1×10^-15) indicating strong protection; APOE4 dose OR 6.00 (5.06–7.12; p=3.4×10^-90) indicating strong risk. - Effect sizes were attenuated in the neuropathologically unconfirmed group: APOE4/4 OR 10.70 (9.12–12.56; p=7.5×10^-186) and APOE2 dose OR 0.64 (0.58–0.72; p=2.2×10^-16), supporting stronger genotype effects when diagnostic misclassification is minimized. - APOE2/2 vs other references in the neuropathologically confirmed group (Table 2): • Versus APOE2/3: OR 0.34 (0.12–0.95; p=0.004), indicating significantly lower risk than the already protective APOE2/3. • Versus APOE4/4: OR 0.004 (0.001–0.041; p=6.0×10^-19), indicating an exceptionally low risk relative to the highest-risk genotype. - Age at onset (neuropathologically confirmed cases): Mean onset age increased across genotypes from APOE4/4 (69.9 ± 6.1 years) to APOE3/3 and to combined APOE2/3 + APOE2/2 (77.8 ± 8.5 years), consistent with protective effects of APOE2 and detrimental effects of APOE4. - Neuropathology severity: APOE2/2 showed lower Braak stage than APOE2/3 (2.26 ± 1.63 vs 3.16 ± 1.69; P=0.05). Versus APOE3/3, APOE2/2 and APOE2/3 had significantly lower CERAD and Braak measures; APOE2/4, APOE3/4, and APOE4/4 had higher scores, before and after adjustment for age and sex. Residual analyses adjusting for CERAD indicated progressively protective effects of APOE2 genotypes and harmful effects of APOE4 genotypes on Braak stage (e.g., versus APOE4/4, β≈–1.32 for APOE2/2; P<0.05). - Other neuropathologies: In the neuropathologically confirmed group, APOE4 allelic dose was associated with higher odds of LBD and HS and with higher odds of CAA and VBI, before and after adjusting for age, sex, and AD status. - Kaplan–Meier survival: Greater APOE2 allelic dose was associated with a higher probability of remaining free from Alzheimer’s dementia to older ages.

Findings support that APOE2 homozygosity confers an exceptionally low likelihood of Alzheimer’s dementia, even lower than APOE2/3, and dramatically lower than APOE4/4. The stronger gene-dose effects observed in neuropathologically confirmed cohorts underscore the value of definitive pathological classification to accurately quantify genetic risk. Protective effects of APOE2 and detrimental effects of APOE4 extend beyond amyloid plaque burden, as residual analyses indicate differential impacts on tau (Braak stage) independent of neuritic plaque severity. The results reinforce APOE biology as a central driver of AD risk and suggest that targeting APOE-related mechanisms could meaningfully affect disease onset and progression. Broader implications include the need for biomarker- or pathology-based case definitions in genetic risk studies and for therapies that modulate APOE function or mimic beneficial APOE2 effects.

APOE2 homozygosity is associated with an exceptionally low risk of Alzheimer’s dementia, with gene-dose effects more pronounced in neuropathologically confirmed samples. APOE genotypes strongly influence odds of disease and neuropathological severity, including tau pathology independent of amyloid plaques. These findings highlight APOE and its pathways as promising therapeutic targets for delaying or preventing Alzheimer’s dementia. Future research should elucidate mechanisms by which APOE variants modulate amyloid, tau, microglial responses, lipid transport, and related pathways, and evaluate interventions (e.g., gene editing, protein modulation) that safely recapitulate protective APOE2-like effects. Studies in diverse racial/ethnic populations and with comprehensive biomarker/neuropathological confirmation are needed.

- Rare APOE2/2 genotype led to relatively small numbers of homozygotes, limiting power for some comparisons. - Lack of neuropathological data for most participants in some contributing cohorts prevents fully disentangling the effects of excluding misclassified cases/controls from potential selection biases associated with brain donation. - Generalizability may be limited beyond clinically defined Alzheimer’s dementia; analyses were largely restricted to non-Hispanic White participants. - Age at onset was missing for some cases and proxied by age at death, which may introduce measurement error. - Some comorbid pathologies (e.g., TDP-43, microinfarcts) were not consistently assessed and were excluded.