CoChat
Sign Up
Cited by 51Alzheimer's disease research and treatmentsDementia and Cognitive Impairment ResearchGenetic Associations and Epidemiology

Association of African Ancestry–Specific APOE Missense Variant R145C With Risk of Alzheimer Disease

Open access
Le Guen, Yann, Raulin, Ana-Caroline, Logue, Mark W., Sherva, Richard, Belloy, Michael E., Eger, Sarah J., Chen, Annabel, Kennedy, Gabriel +13 more·2023·JAMA
DOI 10.1001/jama.2023.0268PMC PMC9945061
Open this paper in CoChat — free accountOpen PDF

Abstract

Importance: Numerous studies have established the association of the common APOE ε2 and APOE ε4 alleles with Alzheimer disease (AD) risk across ancestries. Studies of the interaction of these alleles with other amino acid changes on APOE in non-European ancestries are lacking and may improve ancestry-specific risk prediction. Objective: To determine whether APOE amino acid changes specific to individuals of African ancestry modulate AD risk. Design, Setting, and Participants: Case-control study including 31 929 participants and using a sequenced discovery sample (Alzheimer Disease Sequencing Project; stage 1) followed by 2 microarray imputed data sets derived from the Alzheimer Disease Genetic Consortium (stage 2, internal replication) and the Million Veteran Program (stage 3, external validation). This study combined case-control, family-based, population-based, and longitudinal AD cohorts, which recruited participants (1991-2022) in primarily US-based studies with 1 US/Nigerian study. Across all stages, individuals included in this study were of African ancestry. Exposures: Two APOE missense variants (R145C and R150H) were assessed, stratified by APOE genotype. Main Outcomes and Measures: The primary outcome was AD case-control status, and secondary outcomes included age at AD onset. Results: Stage 1 included 2888 cases (median age, 77 [IQR, 71-83] years; 31.3% male) and 4957 controls (median age, 77 [IQR, 71-83] years; 28.0% male). In stage 2, across multiple cohorts, 1201 cases (median age, 75 [IQR, 69-81] years; 30.8% male) and 2744 controls (median age, 80 [IQR, 75-84] years; 31.4% male) were included. In stage 3, 733 cases (median age, 79.4 [IQR, 73.8-86.5] years; 97.0% male) and 19 406 controls (median age, 71.9 [IQR, 68.4-75.8] years; 94.5% male) were included. In ε3/ε4-stratified analyses of stage 1, R145C was present in 52 individuals with AD (4.8%) and 19 controls (1.5%); R145C was associated with an increased risk of AD (odds ratio [OR], 3.01; 95% CI, 1.87-4.85; P = 6.0 × 10 −6 ) and was associated with a reported younger age at AD onset (β, −5.87 years; 95% CI, −8.35 to −3.4 years; P = 3.4 × 10 −6 ). Association with increased AD risk was replicated in stage 2 (R145C was present in 23 individuals with AD [4.7%] and 21 controls [2.7%]; OR, 2.20; 95% CI, 1.04-4.65; P = .04) and was concordant in stage 3 (R145C was present in 11 individuals with AD [3.8%] and 149 controls [2.7%]; OR, 1.90; 95% CI, 0.99-3.64; P = .051). Association with earlier AD onset was replicated in stage 2 (β, −5.23 years; 95% CI, −9.58 to −0.87 years; P = .02) and stage 3 (β, −10.15 years; 95% CI, −15.66 to −4.64 years; P = 4.0 × 10 −4 ). No significant associations were observed in other APOE strata for R145C or in any APOE strata for R150H. Conclusions and Relevance: In this exploratory analysis, the APOE ε3[R145C] missense variant was associated with an increased risk of AD among individuals of African ancestry with the ε3/ε4 genotype. With additional external validation, these findings may inform AD genetic risk assessment in individuals of African ancestry.

AI snapshot

Written from the paper’s full text.

TL;DR

In African-ancestry individuals with the APOE ε3/ε4 genotype, the ε3[R145C] missense variant was associated with increased Alzheimer disease risk and earlier age at onset. The association replicated internally and was concordant in an external MVP validation, though the external case-control result was just above conventional significance.

Key findings

In stage 1 ε3/ε4-stratified analysis, R145C was present in 52 AD cases (4.8%) vs 19 controls (1.5%) and associated with increased AD risk (OR, 3.01; 95% CI, 1.87-4.85; P = 6.0 × 10−6); stage 2 replicated this (OR, 2.20; 95% CI, 1.04-4.65; P = .04) and stage 3 was concordant (OR, 1.90; 95% CI, 0.99-3.64; P = .051). Among ε3/ε4 AD cases, R145C was associated with younger age at onset in stage 1 (β, −5.87 years; 95% CI, −8.35 to −3.4; P = 3.4 × 10−6), stage 2 (β, −5.23 years; 95% CI, −9.58 to −0.87; P = .02), and stage 3 (β, −10.15 years; 95% CI, −15.66 to −4.64; P = 4.0 × 10−4). No significant associations were observed in other APOE strata for R145C or in any APOE strata for R150H; R150H in ε2/ε3 had OR, 1.08 (95% CI, 0.33-3.55; P = .90). R145C was associated with worsening CDR-SB over time in ε3/ε4-stratified analysis (β, 0.19 increase per year; 95% CI, 0.09-0.29; P = 1.1 × 10−4), and apoE3-R145C eluted at lower NaCl than apoE3 (0.36 M vs 0.49 M; n = 3; P = .026).

Methods

Case-control study including 31,929 participants across three stages: sequenced discovery from the Alzheimer Disease Sequencing Project (stage 1; 2888 cases, 4957 controls), microarray imputed internal replication from the Alzheimer Disease Genetic Consortium (stage 2; 1201 cases, 2744 controls), and external validation in the Million Veteran Program (stage 3; 733 cases, 19,406 controls); all included individuals were of African ancestry. Exposures were APOE missense variants R145C and R150H, stratified by APOE genotype; primary outcome was AD case-control status, with secondary outcomes of age at AD onset and CDR-SB score. Main analyses used APOE-stratified linear mixed-model regression adjusted for sex and 3 genetic principal components, with fixed-effect inverse variance-weighted meta-analysis of stages 1 and 2, plus sensitivity analyses for African-ancestry cutoffs and local ancestry, and a heparin-binding functional assay.

Limitations

Stated in the paper: the main limitation is lack of definitive functional validation, and more detailed mechanistic studies using AD-relevant cell types are needed; the effect of the rare ε3[R145C]/ε3[R145C] genotype remains unanswered and requires larger sample sizes; R145C is known to increase risk for type 3 hyperlipidemia, which could lead to greater vascular pathology and should be evaluated in future studies. Evident from the text: the stage 3 external case-control validation did not meet P < .05 (P = .051), and the R145C-by-ε4 interaction was not significant in stage 3 (P = .98). Evident from the text: stage 3 was overwhelmingly male (97.0% of cases and 94.5% of controls), which may limit generalizability.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1001/jama.2023.0268

Cites

39 of 39 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

12 of 33 cited sources checked · 7 supported · 1 flagged · accuracy not yet measured

  • Findingown data

    In meta-analysis of stages 1 and 2, there was a significant interaction between R145C and APOE ε4 (OR, 2.66; 95% CI, 1.68-4.23; P = 3.4 × 10 -5 ).

    “We for- mally tested this interaction and found a significant associa- tion (OR, 2.66; 95% CI, 1.68-4.23; P = 3.4 × 10 -5 ) (eTable 15 in Supplement 1) in the meta-analysis of stages 1 and 2”
    Quote in the paperSubstantial claim
  • Findingown data

    In meta-analysis of stages 1 and 2, among individuals with APOE ε3/ε4 genotype, R145C was associated with increased Alzheimer disease risk (OR 2.75, 95% CI 1.84-4.11, P=8.3×10^-7).

    “Meta-analysis | 3658 | 115 | 2.75 (1.84-4.11) | 8.3 × 10 -7 1484 | 72 | -5.72 (-7.87 to -3.56) | 2.0 × 10 -7 | |”
    Quote in the paperSubstantial claim
  • Findingown data

    In the ADSP discovery stage, among individuals with APOE ε3/ε4 genotype, R145C was associated with earlier age at onset of Alzheimer disease (β = -5.87, 95% CI -8.35 to -3.4, P=3.4×10^-6).

    “Discovery | 2381 | 71 | 3.01 (1.87-4.85) | 6.0 × 10 -6 1063 | 51 | -5.87 (-8.35 to -3.4) | 3.4 × 10 -6 | |”
    Quote in the paperSubstantial claim
  • Findingown data

    In the stage 1 discovery sample, among individuals with APOE ε3/ε4 and AD, R145C was associated with a younger age at AD onset (β, -5.87 years; 95% CI, -8.35 to -3.4 years; P = 3.4 × 10 -6 ).

    “and was associated with a reported younger age at AD onset (β, -5.87 years; 95% CI, -8.35 to -3.4 years; P = 3.4 × 10 -6 )”
    Quote in the paperSubstantial claim
  • Findingown data

    In Japanese ancestry, compared with APOE ε3/ε3, the APOE ε4/ε4 genotype was not significantly associated with Alzheimer disease risk (OR 2.2, 95% CI 0.7-6.7).

    “0.6 (0.3-1.3) | 1 [Reference] | 2.2 (1.3-3.4) | 2.2 (0.7-6.7)”
    Quote in the paperNot a substantial claim
  • May combine two claims. 19
  • Classified as probably not a substantial claim. 17
  • The quote only partly states this claim. 3

  • May combine two claims.

    Premisecitation

    R145C is always in phase with APOE ε3 and R150H is always in phase with APOE ε2.

    “Prior studies [7][8][9] and Alzheimer Disease Sequencing Project (ADSP) data have shown that R145C is always in phase with ε3 and R150H is always in phase with ε2.”
    Quote in the paperNot a substantial claim1 of 3 sources checked
    • cites [7] Abou Ziki, 2014, Prevalence of the apolipoprotein E Arg145Cys dyslipidemia at-risk polymorphism in African-derived populations., Am J Cardiol source

      The Results (page not provided) say that most African-American R145C variants were on the E3 allele, but identify one on an E4 allele. That counterexample contradicts the claim that R145C is always in phase with ε3. The retrieved passages do not address R150H phase. (LLM, conf=0.89)

      Checked by a reasoning model — may be wrong.

      Experimental — accuracy not yet measured
    • cites [8] Rall, 1982, Structural basis for receptor binding heterogeneity of apolipoprotein E from type III hyperlipoproteinemic subjects., Proc Natl Acad Sci U S A source
    • cites [9] de Villiers, 1997, The apolipoprotein E2 (Arg145Cys) mutation causes autosomal dominant type III hyperlipoproteinemia with incomplete penetrance., Arterioscler Thromb Vasc Biol source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    The MVP cohort used in this study was assembled for an African-ancestry dementia genome-wide association study described by Sherva et al.

    “The MVP cohort used herein was assembled for an African-ancestry dementia genome-wide association study described by Sherva et al. 17”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [17] Sherva, 2022, African ancestry GWAS of dementia in a large military cohort identifies significant risk loci. source

  • May combine two claims.

    Study design and conductcitation

    Case-control analyses were not adjusted for age because controls were older than cases in some APOE ε3 genotype strata, and correcting for age when cases are younger than controls leads the model to incorrectly infer the age effect on AD risk, resulting in statistical power loss.

    “Case-control analyses were not adjusted for age given that controls were older than cases in some APOE ε3 genotype strata (Table 2).Correcting for age when cases are younger than controls leads the model to incorrectly infer the age effect on AD risk, resulting in statistical power loss.20”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [20] Le Guen, 2021, A novel age-informed approach for genetic association analysis in Alzheimer’s disease., Alzheimers Res Ther source

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductcitation

    All statistical analyses were adjusted for sex and 3 genetic principal components estimated with the PC-air method, and linear mixed-model analyses were additionally covaried by a sparse genetic relationship matrix estimated with the PC-Relate method.

    “All statistical analyses were adjusted for sex and 3 genetic principal components estimated with the PC-air method 29 implemented in GENESIS.Linear mixed-model analyses were additionally covaried by a sparse genetic relationship matrix estimated with the PC-Relate method 30 implemented in GENESIS.”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites [29] Conomos, 2015, Robust inference of population structure for ancestry prediction and correction of stratification in the presence of relatedness., Genet Epidemiol source
    • cites [30] Conomos, 2016, Genetic diversity and association studies in US Hispanic/Latino populations: applications in the Hispanic Community Health Study/Study of Latinos., Am J Hum Genet source

      Split into 2 atoms — 1 confirmed, 1 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

      Experimental — accuracy not yet measured

  • Classified as probably not a substantial claim.

    Premisecitation

    When examining variants at the APOE locus, APOE-stratified analyses may provide better statistical power than standard APOE-adjusted models.

    “Recent work has suggested that when examining variants at the APOE locus, APOE-stratified analyses may provide better statistical power than standard APOE-adjusted models.26”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [26] Jun, 2016, A novel Alzheimer disease locus located near the gene encoding tau protein., Mol Psychiatry source

      The Discussion (page unavailable) says APOE-stratified analysis revealed an association at the chromosome 17q21.31/MAPT region that was not genome-wide significant without stratification. This supports the possibility that stratification can reveal otherwise obscured associations, but the paper does not establish greater statistical power for variants at the APOE locus or compare stratified analyses with standard APOE-adjusted models. (LLM, conf=0.84)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    V236E and R251G have been associated with reduced AD risk.

    “V236E and R251G have been associated with reduced AD risk.4,5”
    Quote in the paperNot a substantial claim2 of 2 sources checked
    • cites [4] Medway, 2014, ApoE variant p.V236E is associated with markedly reduced risk of Alzheimer’s disease., Mol Neurodegener source
    • cites [5] Le Guen, 2022, Association of rare APOE missense variants V236E and R251G with risk of Alzheimer disease., JAMA Neurol source

  • Classified as probably not a substantial claim.

    Premisecitation

    Heparan sulfate proteoglycans have been shown to regulate cellular uptake of tau.

    “since heparan sulfate proteoglycans have been shown to regulate cellular uptake of tau.33”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [33] Rauch, 2018, Tau internalization is regulated by 6-O sulfation on heparan sulfate proteoglycans (HSPGs)., Sci Rep source

  • May combine two claims.

    Premisecitation

    Local ancestry analysis has confirmed that R145C is always on an African haplotype and thus is seen only in individuals with African admixture.

    “Local ancestry analysis 10 has confirmed that this variant is always on an African haplotype (eFigure 1 in Supplement 1) and thus is seen only in individuals with African admixture.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [10] Karczewski, 2020, The mutational constraint spectrum quantified from variation in 141,456 humans., Nature source

      The population comparison on p. 439 concerns gene-level constraint, while the Online content describes sample-level ancestry inference. Neither addresses R145C or whether it occurs exclusively on African haplotypes. Searches for “R145C” and “local ancestry African haplotype admixture” found no passage establishing the claim. (LLM, conf=0.74)

      Experimental — accuracy not yet measured

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    MVP analyses were adjusted for sex, 3 principal components accounting for population stratification, and age at last examination.

    “These analyses were adjusted for sex, 3 principal components accounting for population stratification, and age at last examination, since the MVP is a population-based data set.20”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [20] Le Guen, 2021, A novel age-informed approach for genetic association analysis in Alzheimer’s disease., Alzheimers Res Ther source

      The Methods section ('Statistics, association models, and AD-age score'; page unavailable) reports adjustment for sex and population-structure principal components, but specifies 10 components, not 3. The Methods ('Participants'; page unavailable) mentions age at last examination as a possible age measure for controls, not as an adjustment variable. These passages do not establish the stated adjustments for the MVP analyses. (LLM, conf=0.85)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Two recent studies have suggested that the minimum clinically important difference over 1 year is in the range of 1 to 2 points.

    “Two recent studies have suggested that the minimum clinically important difference over 1 year is in the range of 1 to 2 points.24,25”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites [24] Andrews, 2019, Disease severity and minimal clinically important differences in clinical outcome assessments for Alzheimer’s disease clinical trials., Alzheimers Dement (N Y) source
    • cites [25] Wessels, 2022, Integrated Alzheimer’s Disease Rating Scale: clinically meaningful change estimates., Alzheimers Dement (N Y) source

  • May combine two claims.

    Premisecitation

    More than 4% of individuals of African ancestry carry R145C, while it is very rare in people of European ancestry.

    “More than 4% of individuals of African ancestry carry R145C, 7 while it is very rare in people of European ancestry (eTable 1 in Supplement 1).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [7] Abou Ziki, 2014, Prevalence of the apolipoprotein E Arg145Cys dyslipidemia at-risk polymorphism in African-derived populations., Am J Cardiol source

  • The quote only partly states this claim.

    Findingown data

    In Japanese ancestry, compared with APOE ε3/ε3, the APOE ε2/ε3 genotype was not significantly associated with Alzheimer disease risk (OR 0.6, 95% CI 0.3-1.3).

    “0.6 (0.3-1.3) | 1 [Reference] | 2.2 (1.3-3.4) | 2.2 (0.7-6.7)”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    In a non-APOE-stratified analysis, the association of R145C with increased AD risk was weaker, while nominally significant.

    “In a non-APOE-stratified analysis, the association of R145C with increased AD risk was weaker, while nominally significant, and it was driven by the association in the ε3/ε4 subgroup.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    In mega-analysis of stages 1 and 2, the ε3/ε4 association with AD risk reached genome-wide significance (R145C was present in 75 cases [4.7%] and 40 controls [1.9%]; OR, 2.93; 95% CI, 1.99-4.31; P = 4.80 × 10 -8 ).

    “Results in the mega-analysis of stages 1 and 2 were similar com- pared with the meta-analysis design, and in the ε3/ε4 group, the association with AD risk reached genome-wide signifi- cance (R145C was present in 75 cases [4.7%] and 40 controls [1.9%]; OR, 2.93; 95% CI, 1.99-4.31; P = 4.80 × 10 -8 ) (eTable 12 in Supplement 1).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Restricting ε3/ε4-stratified analyses to individuals with African local ancestry at both APOE haplotypes led to similar point estimates (OR, 3.10; 95% CI, 1.68-5.70; P = 2.80 × 10 -4 ) with larger P values due to reduced sample sizes.

    “Additionally, restrict- ing our ε3/ε4-stratified analyses to individuals with African lo- cal ancestry at both APOE haplotypes led to similar point es- timates (OR, 3.10; 95% CI, 1.68-5.70; P = 2.80 × 10 -4 ) with larger P values, compared with the unrestricted discovery sample, due to reduced sample sizes (eTable 11 in Supplement 1).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    In sensitivity analyses, results remained unchanged when selecting admixed-ancestry individuals with at least 45% African ancestry or when restricting the analysis to African-ancestry individuals using a 75% cutoff.

    “Briefly, the results remained unchanged when selecting admixed- ancestry individuals with at least 45% African ancestry or when restricting the analysis to African-ancestry individuals (75% cutoff).”
    Quote in the paperNot a substantial claim

  • The quote only partly states this claim.

    Findingown data

    In stage 2 internal replication, among AD cases with APOE ε3/ε4, R145C was associated with an earlier age at AD onset (β, -5.23 years; 95% CI, -9.58 to -0.87 years; P = .02).

    “Association with earlier AD onset was replicated in stage 2 (β, -5.23 years; 95% CI, -9.58 to -0.87 years; P = .02)”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    In stage 1 discovery, using a 75% African ancestry cutoff yielded an OR of 3.40 (95% CI, 1.95-5.90; P = 1.50 × 10 -5 ), compared with an OR of 3.01 using a 15% cutoff.

    “For example, using an ancestry cutoff at 75% in the stage 1 discovery yielded an OR of 3.40 (95% CI, 1.95-5.90; P = 1.50 × 10 -5 ), compared with an OR of 3.01 using a cutoff of 15%.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    In the discovery sample, apart from the ε2 and ε4 alleles, the canonical APOE transcript harbored 2 variants with minor allele count above 10: R145C with a minor allele frequency of approximately 2% and R150H with a minor allele frequency of approximately 0.1%.

    “Apart from the ε2 and ε4 alleles, the canonical APOE transcript harbored 2 variants with minor allele count above 10 in the discovery sample used in this study.One is common, Arg145Cys (R145C), with a minor allele frequency of approximately 2%, and one is rare, Arg150His (R150H), with a minor allele frequency of approximately 0.1% (eTable 1 in Supplement 1).”
    Quote in the paperSubstantial claim

  • The quote only partly states this claim.

    Findingown data

    In European ancestry, compared with APOE ε3/ε3, the APOE ε2/ε3 genotype was not significantly associated with Alzheimer disease risk (OR 0.9, 95% CI 0.4-2.5).

    “| | | Odds ratio | (95% CI) | | | 0.9 (0.4-2.5) | 1 [Reference] | 5.6 (3.9-8.0) | 33.1 | (13.6-80.5) |”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    In stage 3 external validation, R145C was associated with AD risk in a concordant direction but the association was not statistically significant (OR, 1.90; 95% CI, 0.99-3.64; P = .051).

    “OR, 1.90; 95% CI, 0.99-3.64;P = .051)”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Kunkle et al reported a similar OR in their nonstratified analysis (OR, 1.38; 95% CI, 0.98-1.94; P=.056).

    “Of note, Kunkle et al 15 reported a similar OR in their nonstratified analysis (OR, 1.38; 95% CI, 0.98-1.94;P = .056)conducted on a slightly smaller subset of samples than used here.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Kunkle et al 15

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    R145C is known to increase risk for type 3 hyperlipidemia.

    “Third, R145C is known to increase risk for type 3 hyperlipidemia, [7][8][9]39 which could lead to a greater degree of vascular pathology and should be evaluated in future studies.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [7][8][9]39

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductnone given

    Drs Le Guen and Greicius had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

    “Drs Le Guen and Greicius had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductnone given

    Quality control followed standard practices and was performed per cohort-genotyping platform pair.

    “Quality control followed standard practices and was performed per cohort-genotyping platform pair and is described in eAppendix 2 in Supplement 1.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Kunkle et al did not test the interaction of R145C with ε4 for association with AD status.

    “Kunkle et al 15 did not, however, test the interaction of R145C with ε4 for association with AD status.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Kunkle et al 15

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    Discovery analyses were considered significant if they reached a Bonferroni-corrected P value threshold of .0125, accounting for the analysis of R145C in 3 APOE strata and R150H in ε2/ε3.

    “The discovery analyses were considered significant if they reached a Bonferroni-corrected P value threshold of .0125(≈ .05/4),which accounts for the analysis of R145C in 3 APOE strata (ε2/ε3, ε3/ε3, and ε3/ε4) and R150H in ε2/ε3.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    The lack of definitive functional validation stands as the main limitation of this study.

    “First, the lack of definitive functional validation stands as the main limitation of this study.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    Sensitivity analyses using the 45% and 75% ancestry cutoffs suggest that the significance of the results is independent of the African-ancestry cutoffs and not due to population stratification.

    “Sensitivity analyses using the 45% and 75% ancestry cutoffs led to the same conclusion (eFig- ure 7 in Supplement 1), suggesting that the significance of the results is independent of the African-ancestry cutoffs and not due to population stratification.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    Replications were considered significant at P < .05 with a concordant direction.

    “Only the R145C ε3/ε4-stratified analyses were significant in the discovery; thus, replications were considered significant at P < .05 with a concordant direction.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    By inclusion criteria, all participants had no missing data, having both AD diagnosis and genotype information available.

    “By inclusion criteria, all participants had no missing data, ie, they had both AD diagnosis (case or control) and genotype information available.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    ApoE3-R145C does not lead to a gain of deleterious function since it does not increase AD risk in the absence of ε4.

    “This suggests that apoE3-R145C does not lead to a gain of deleterious function since it does not increase AD risk in the absence of ε4.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductown data

    Individuals with an imputation quality R2 greater than 0.8 at rs769455 were retained.

    “We retained individuals with an R 2 greater than 0.8 at rs769455 (eTable 6 in Supplement 1).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductown data

    This study's main finding is supported by 2 independent replications and concordant secondary analyses.

    “However, while further functional assays testing the hypotheses proposed above are needed, this study's main finding is supported by 2 independent replications and concordant secondary analyses.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductown data

    The findings persisted in sensitivity analyses examining different degrees of African ancestry and when restricting to individuals with African local ancestry at the APOE locus, suggesting that these findings were not confounded by differences in local ancestry.

    “The findings persisted in sensitivity analyses examining different degrees of African ancestry and when restricting the analysis to individuals with African local ancestry at the APOE locus, suggesting that these findings were not confounded by differences in local ancestry and that R145C is the causal variant.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    The AD risk associated with common APOE alleles ε4 and ε2 is well established in European-ancestry individuals but varies considerably across other ancestries.

    “The AD risk associated with common APOE alleles ε4 1 and ε2 2 is well established in European-ancestry individuals but varies considerably across other ancestries (Table 1). 3”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites [1] Corder, 1993, Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer’s disease in late onset families., Science source
    • cites [2] Corder, 1994, Protective effect of apolipoprotein E type 2 allele for late onset Alzheimer disease., Nat Genet source
    • cites [3] Farrer, 1997, Effects of age, sex, and ethnicity on the association between apolipoprotein E genotype and Alzheimer disease: a meta-analysis., JAMA source
36 of 36

  • The paper relies on this without support.

    Additional APOE missense variants have been understudied because they are present mainly in African-ancestry individuals, who are underrepresented in AD genetics studies.

    “Additional APOE missense variants have been understudied because they are present mainly in African-ancestry individuals, who are underrepresented in AD genetics studies.”

  • The paper relies on this without support.

    R145C is located within apoE's receptor-binding region at amino acid residues 136 to 150 in the N-terminal domain.

    “The R145C variant is located within apoE's receptorbinding region at amino acid residues 136 to 150 in the N-terminal domain.”
2 of 2

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Specific APOE rare missense variants (p.L46P, p.R154S, p.R163C, p.V254E, p.R269G) have been reported in Western populations and suggest potential resistance to amyloid-β accumulation, phosphorylated tau pathology, or the effects of the APOE ε4 allele.

    “For example, RMVs such as p.L46P (p.L28P; rs769452 [Pittsburgh]), 17 p.R154S (p.R136S; rs121918393 [Christchurch]), 18 p.R163C (p.R145C; rs769455), 19 p.V254E (p.V236E; rs199768005 [Jacksonville]), 20 and p.R269G (p.R251G; rs267606661) 21 have been reported in Western populations, suggesting potential resistance to amyloid-β (Aβ) accumulation, phosphorylated tau pathology, or the effects of the AP”

    From Association of rare <i>APOE</i> missense variants with Alzheimer's disease in the Japanese population (2025) · cites this paper as [19]

    Split into 2 atoms — 1 confirmed, 0 partially supported, 1 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

    Checked by a reasoning model — may be wrong.

    Experimental — accuracy not yet measured

  • APOE3-R145C is associated with increased risk of AD compared with ApoE3.

    “APOE3-R145C has been shown to increase the risk of AD 38 .”

    From Utilizing a cell culture based novel cellular thermal shift assay to understand the isoform-dependent thermal stability of ApoE variants (2026) · cites this paper as [38]

    The Results and Discussion report increased Alzheimer disease risk associated with APOE3-R145C in African-ancestry individuals with the ε3/ε4 genotype. The citing sentence omits that restriction: the paper reports no significant increase among ε3/ε3 carriers, so it does not support a general increase in risk compared with ApoE3. (LLM, conf=0.91)

    Experimental — accuracy not yet measured

  • APOE4-R251G decreases the risk of ApoE4 to be similar to that of ApoE3.

    “ApoE4-R251G has been shown to decrease the risk of ApoE4 to be similar to that of ApoE3 37,38”

    From Utilizing a cell culture based novel cellular thermal shift assay to understand the isoform-dependent thermal stability of ApoE variants (2026) · cites this paper as [38]

    The introductory passage labeled Abstract (page unavailable) says R251G has been associated with reduced Alzheimer disease risk. It does not say the reduction occurs specifically on an ApoE4 background or brings risk to the level of ApoE3; the paper’s reported analysis concerns R145C instead. (LLM, conf=0.86)

    Experimental — accuracy not yet measured

  • Among individuals of African ancestry who have one copy of e3 and one of e4, those with a particular variant of e3 called R145C are at increased risk of developing Alzheimer's dementia compared with those who do not have this variant.

    “Among individuals of African ancestry who have one copy of e3 and one of e4, those with a particular variant of e3 called R145C are at increased risk of developing Alzheimer's dementia compared with those who do not have this variant.”

    From 2024 Alzheimer's disease facts and figures (2024) · cites this paper as [95]

  • In African ancestry populations, the ε3-linked R145C variant increases AD risk in a genotype-specific manner.

    “In African ancestry populations, the ε3-linked R145C variant increases AD risk in a genotype-specific manner 4 .”

    From Association of Rare <i>APOE</i> Missense Variants R189C and W276C With Risk of Alzheimer Disease (2026) · cites this paper as [4]

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Stage 3 ε2/ε2 case percentage is too small

    Minor

    The table reports 1 ε2/ε2 case among 733 stage 3 cases. That is 0.136%, not 0.013%.

    Reported 0.013 · recomputed 0.136426

  • An IQR is reported for a single stage 3 ε2/ε2 case

    Minor

    The table lists only 1 stage 3 case with the ε2/ε2 genotype, but gives that group's age an IQR with distinct endpoints. A single observation cannot have an IQR different from its median.

2 of 2

Cites

Flagged first, then by citations

4 of the 43 references couldn’t be matched to a known paper.

Cited by

Sorted by citations

Related

Authors
Le Guen, Y., Raulin, A.-C., Logue, M. W., Sherva, R., Belloy, M. E., Eger, S. J., Chen, A., Kennedy, G., Kuchenbecker, L., O’Leary, J. P., Zhang, R., Merritt, V. C., Panizzon, M. S., Hauger, R. L., Gaziano, J. M., Bu, G., Thornton, T. A., Farrer, L. A., Napolioni, V., He, Z., & Greicius, M. D.
Affiliations
  • Department of Neurology and Neurological Sciences, Stanford University, Stanford, California
  • Mayo Clinic in Florida (US)
  • Boston University (US)
  • Stanford University (US)
  • VA Boston Healthcare System (US)
  • University of California San Diego (US)
  • Brigham and Women's Hospital (US)
  • University of Washington (US)
  • Università di Camerino (IT)
ORCID
Venue
JAMA
Publisher
American Medical Association (AMA)
Published
2023-02-21
Volume
329
Issue
7
Pages
551
ISSN
  • 0002-9955 (ISSN-L)
  • 0098-7484
DOI
10.1001/jama.2023.0268
PMC
PMC9945061
OpenAlex
W4321446458
Semantic Scholar
7aeef1f76753971a31fe237e2f60a7424385f55b
Cited by
50
References
43
Citations by year
2026: 11 · 2025: 17 · 2024: 12 · 2023: 10
FWCI
8.88
Type
journal article
Language
en
Open access
Locations
  • JAMA (publishedVersion)
  • PubMed Central (publishedVersion)
  • PubMed (publishedVersion)
  • Unicam Scientific Publications (University of Camerino) (submittedVersion)
  • Rare & Special e-Zone (The Hong Kong University of Science and Technology) (submittedVersion)
Funders
  • U.S. Department of Veterans Affairs (I01 BX004192, I01 CX001727)
  • Alzheimer's Association (AARF-20-683984)
  • National Institute on Aging (U01AG058654, R01AG060747, R01AG048927, U01AG062602, RF1AG057519, R01AG066206, U19AG068753, P30AG066515, U24AG021886)
Topics
  • Health Sciences › Medicine › Physiology › Alzheimer's disease research and treatments
  • Health Sciences › Medicine › Psychiatry and Mental health › Dementia and Cognitive Impairment Research
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Genetics › Genetic Associations and Epidemiology
Keywords
Alzheimer's disease, apolipoprotein E, missense mutations, African ancestry, genetic risk, age of onset, APOE ε2, APOE ε4, Million Veteran Program, genotype imputation, family-based association, longitudinal cohort
MeSH terms
Aged, Aged, 80 and over, Alleles, Alzheimer Disease, Apolipoproteins E, Female, Genotype, Humans, Male, Risk Factors, Case-Control Studies, Mutation, Missense, Black People, Apolipoprotein E4

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-31.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-31.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review