APOE is the strongest common genetic influence on late-onset Alzheimer's disease. Its e4 form raises risk and e2 lowers it, but APOE genotype predicts probability rather than destiny.
Solid lines are connections this site curates. Dashed lines mean the two ends share a research paper — worth knowing, and not a claim that one explains the other.
Alzheimer's disease is the most common cause of dementia. It develops gradually, typically beginning with difficulty forming new memories, and progresses over years to affect language, judgment, orientation and eventually the ability to manage daily life. In the brain, two abnormal protein deposits accumulate: amyloid-beta plaques between nerve cells and tangles of a protein called tau inside them, accompanied by inflammation and the loss of neurons and their connections.
The great majority of Alzheimer's disease — around 95% — is the late-onset form, appearing after roughly age 65. A small minority is early-onset, and only a fraction of that is caused by rare single-gene mutations passed directly down families. For the common late-onset form, genetics works differently: many genes each nudge risk up or down, and one gene stands out above all the others.
The APOE gene makes apolipoprotein E, a protein that shuttles cholesterol and other fats around the body and, in the brain, helps supply neurons with lipids and participates in clearing amyloid away. Almost everyone carries two copies of APOE, one from each parent, and each copy comes in one of three common versions, called e2, e3 and e4. Your combination — for example e3/e3 or e3/e4 — is your APOE genotype.
The e2/e3/e4 versions are not one variant but a pair working together. Two positions in the gene — known by their catalog numbers rs429358 and rs7412 — are read as a combination, and it is the pairing of the two that tells you which version a chromosome carries. That is why APOE is reported as a genotype rather than as a single yes-or-no result.
A 2021 study took a different approach to this same region: instead of a clinical Alzheimer's diagnosis, it measured amyloid plaque deposition directly in the brains of roughly 1,000 people using PiB-PET imaging, then ran the first genome-wide search for genetic influences on that specific measurement. As expected, the APOE region dominated the results, with the e4-defining position the single strongest signal in the whole genome. The less expected part: after statistically accounting for e4 itself, 14 more positions in the same broad region remained significantly associated with amyloid deposition, and were not simply riding along with e4 through linkage. This page holds 11 of those variants, spanning APOE, APOC1, TOMM40 and PVRL2 — genes that sit physically close together on chromosome 19 and are often inherited as a block. What this means in practice: the e2/e3/e4 genotype captures the single largest piece of this region's influence on amyloid, but not all of it.
This is the part most worth being careful about. Carrying e4 increases probability, not certainty. Plenty of people with one or even two copies of e4 live long lives without developing dementia, and a large share of people who do develop Alzheimer's disease carry no e4 at all — roughly 42% of them, by one widely cited estimate. So e4 does not diagnose anyone, and its absence does not offer protection or reassurance.
Risk also rises with the number of copies: one copy of e4 carries a moderate increase over the e3/e3 baseline, and two copies a considerably larger one. Recent research has argued that having two copies of e4 behaves less like an ordinary risk factor and more like its own genetically determined form of the disease, because nearly all such individuals eventually show Alzheimer's biological changes on scans and in spinal fluid — though whether and when those changes become symptoms varies from person to person, and this framing is still debated among specialists.
For healthy people with no symptoms, professional genetics guidance has long been that APOE genotyping has little predictive value in practice — it is neither sensitive enough nor specific enough to tell an individual what will happen to them. That said, the picture has shifted somewhat: newer anti-amyloid antibody treatments for people who already have symptoms have made APOE status relevant to weighing the safety of those medicines, which is described in the clinical section below.
Because an APOE result can be genuinely distressing and cannot be reversed once known, people considering testing — or reacting to a result from a consumer test — are usually best served by talking it through with a genetic counselor or physician. Risk is also not fixed by genes alone: cardiovascular health, hearing, sleep, physical activity, education and other factors all influence dementia risk, and unlike genotype they can change.
Positions joined since this page was written
What this is The text above discusses the variants this page was written around. Since then the catalogue has joined 31 more positions to it, by shared trait or shared paper. They are listed here by the paper each came from; the text does not describe them, and each variant page carries that study's own record.
Bellenguez C et al. 2022, Nature genetics rs6605556 (near HLA-DQA1), rs7767350 (CD2AP), rs6586028 (TSPAN14), rs7225151 (ZNF594-DT), rs74685827 (SORL1), rs60755019 (near TREML2), rs3848143 (SNX1), rs10947943 (OARD1), rs6489896 (TPCN1) — PMID:35379992
Willett JDS et al. 2025, Alzheimer's & dementia : the journal of the Alzheimer's Association rs446037 (APOE), rs435380 (APOE), rs72654461 (APOE), rs72654463 (APOE), rs8105340 (NECTIN2), rs166907 (LNCOB1), rs115908094 (TOMM40), rs35114168 (BIN1) — PMID:39998322
Chung J et al. 2023, Alzheimer's & dementia : the journal of the Alzheimer's Association rs34710518 (LEF1-AS1), rs7155666 (DIO2-AS1), rs896559 (PRICKLE1), rs116668122 (SYNPR) — PMID:35770850
Lake J et al. 2023, Molecular psychiatry rs117807585 (SORL1), rs9394766 (TREML2), rs74825460 (FERMT2) — PMID:37198259
Rajabli F et al. 2025, Genome biology rs111486601 (LINC01234), rs1898895 (near RNU6-560P) — PMID:40676597
Ge YJ et al. 2024, Alzheimer's & dementia : the journal of the Alzheimer's Association rs3804082 (OSBP2), rs6682554 (near KIAA2013) — PMID:39023044
EADB 2026, Nature genetics rs58124010 (ZNF594-DT) — PMID:42237039
Wang H et al. 2021, Frontiers in aging neuroscience rs6431220 (near BIN1) — PMID:34122051
Willett JDS et al. 2026, Alzheimer's & dementia : the journal of the Alzheimer's Association rs182804071 (ZFYVE1) — PMID:41755704
2026-08-31 · Washington University in St. Louis (WashU Medicine), reporting on Belloy et al.
Scientists identify genes that could reduce APOE4 Alzheimer's disease risk
Analysis of nearly 450,000 genomes (European, African and Japanese ancestries) found 44 genes linked to why some APOE4 carriers never develop Alzheimer's while others do, with the strongest signal in oligodendrocytes (the cells that produce myelin) -- pointing to drug targets aimed at protecting myelin specifically in APOE4 carriers, this page's own rs429358/rs7412 genotype. Early-stage research, not a treatment.
2026-07-14 · Eisai / Biogen (FDA supplemental approval)
The FDA approved a once-weekly at-home autoinjector (500mg, two 250mg shots) as a starting dose for lecanemab -- previously patients had to complete 18 months of IV infusion before any at-home dosing was possible. This page's own blog post on lecanemab explains why APOE4 genotype (this page's rs429358/rs7412) matters before starting: genotype-stratified brain-swelling risk was the basis for the FDA's pre-treatment testing recommendation.
2026-09-09 · Xu Y, et al. Plasma phosphorylated tau 217 concentrations, APOE genotype, and timing of cognitive impairment in individuals across diverse racial and ethnic groups: a pooled analysis of prospective cohort studies. Lancet Neurology. 2026. DOI:10.1016/S1474-4422(26)00313-3
A Columbia University-led pooled analysis of prospective cohort studies, covering 8,582 people from different racial and ethnic backgrounds who had no cognitive symptoms at the start, found that combining a blood test for p-tau217 (a biomarker of Alzheimer's-related brain changes) with APOE genotype can forecast when symptoms will begin. Once p-tau217 levels became elevated, people carrying one or more copies of the high-risk APOE4 variant -- the same genotype this catalogue's own condition page is about -- went on to develop symptoms in about 3 to 4 years, versus 5 to 6 years for people with other APOE genotypes. Notably, this timing difference held consistently across the different ethnic groups studied, addressing a real prior gap in how well APOE-based prediction generalizes beyond the mostly European-ancestry cohorts most Alzheimer's genetic research has relied on. This is a predictive-timing finding built on the existing, already-catalogued APOE4 genotype, not a new variant -- no new variant page was needed.
2026-06-03 · Bellenguez C, Lambert JC, et al. (Institut Pasteur de Lille, 508 co-authors), Nature Genetics
The largest Alzheimer's GWAS meta-analysis to date (128,681 cases, 849,833 controls) found 91 genetic risk loci in total -- 16 newly discovered, plus 25 new signals across 16 previously known loci, confirming those regions are more genetically complex than earlier studies suggested. PMID:42237039.
2026-09-24 · Schuldt BR, et al. (Blanchard lab, Icahn School of Medicine at Mount Sinai), Cell
A pericyte-to-myofibroblast transition links APOE4 to cerebrovascular degeneration
A single-cell atlas of human brain vasculature shows how APOE4 -- the strongest genetic Alzheimer's risk factor -- damages blood vessels: in APOE4 carriers, pericytes (the cells that support vessel walls) were significantly reduced and a scar-forming, myofibroblast-like cell population appeared in their place. APOE4 pericytes transition into myofibroblasts that secrete fibronectin, which promotes vascular amyloid accumulation; TGF-beta signaling drives the transition, and blocking TGF-beta restored pericyte coverage and brought vascular fibrosis and amyloid back to APOE3 levels in human tissue and mouse models -- a targetable mechanism, per the authors. PMID:42785300.
2026-06-03 · EADB, et al., Nature Genetics 2026, PMID:42237039
Nine research consortia pooled their European-ancestry studies: 128,681 cases or proxy cases of Alzheimer's disease and related dementias and 849,833 controls. The analysis identified 91 loci, 16 of them new; 56 were detected specifically in clinically diagnosed Alzheimer's disease, and a further 18 are listed as needing outside validation. A polygenic score built from the loci other than APOE went with Alzheimer's pathology rather than other kinds: people in the top tenth of the score had about twice the risk of advanced tangle and plaque pathology at death, compared with people in the middle of the range. 69 of the study's positions are already on this site, 59 of them on the Alzheimer's disease page, including rs429358 (APOE), rs7232 (MS4A6A) and rs17462136 (CASS4). PMID:42237039.
Gene and allele definition. APOE (OMIM 107741) on 19q13.32 encodes apolipoprotein E. The e2/e3/e4 alleles are defined by the haplotype of two coding variants in exon 4, reported against NM_000041.4 / NP_000032.1: rs429358, c.388T>C (p.Cys130Arg), and rs7412, c.526C>T (p.Arg176Cys). Legacy literature numbers these residues 112 and 158, counting from the mature protein after the 18-residue signal peptide. Haplotypes: e2 = rs429358-T + rs7412-T (Cys130/Cys176); e3 = rs429358-T + rs7412-C (Cys130/Arg176); e4 = rs429358-C + rs7412-C (Arg130/Arg176). The rare e1 haplotype (rs429358-C + rs7412-T) exists but is not routinely reported. Because the two SNPs are in cis on the same chromosome, a genotype such as rs429358 C/T with rs7412 C/T is ambiguous between e2/e4 and e1/e3 without phasing; laboratories generally resolve this by rarity assumption or direct haplotyping.
Effect estimates. The classical meta-analysis (Farrer et al., JAMA 1997; PMID 9343467) reported odds ratios for AD relative to e3/e3 in participants of European ancestry of 3.2 (95% CI 2.8–3.8) for e3/e4, 14.9 (10.8–20.6) for e4/e4, 2.6 (1.6–4.0) for e2/e4, and 0.6 (0.5–0.8) for e2/e3. GeneReviews summarizes this as roughly a threefold effect for heterozygotes and fifteenfold for homozygotes, corresponding to approximately a 10%–20% chance of AD by age 75 for an e4 heterozygote and 25%–35% for an e4 homozygote. Effect size is ancestry-dependent — attenuated in individuals of African ancestry and in some East Asian and Hispanic/Latino cohorts — and attenuates with advancing age, so genotype-specific estimates derived from European-ancestry cohorts should not be applied uncritically across populations.
Mechanism. ApoE isoforms differ in receptor binding, lipidation and stability. ApoE4 is associated with impaired amyloid-beta clearance and enhanced aggregation, altered lipid transport, increased blood-brain-barrier and cerebrovascular pathology including cerebral amyloid angiopathy, isoform-specific effects on tau-mediated neurodegeneration and microglial state, and a domain-interaction/proteolytic-susceptibility profile distinct from ApoE3. ApoE2 shows markedly reduced LDL-receptor binding, which underlies both its apparent protection in AD and its role in familial dysbetalipoproteinemia (hyperlipoproteinemia type III, OMIM 617347), a condition for which e2/e2 homozygosity is usually necessary but not sufficient.
e4 homozygosity as a distinct entity. Fortea et al. (Nature Medicine 2024; PMID 38710950) analyzed 3,297 brain donors and 10,039 clinically characterized individuals and reported that APOE4 homozygotes show near-complete penetrance of AD neuropathology, biomarker abnormality beginning around age 55, abnormal CSF amyloid in nearly all by 65 with amyloid PET positivity in about three-quarters, and a predictable symptom-onset distribution (median around 65) resembling autosomal dominant AD and Down syndrome-associated AD. The authors argue e4/e4 constitutes a genetic form of AD rather than merely a risk state; the framing has been contested, principally over whether biomarker penetrance should be equated with clinical penetrance.
Clinical utility. Genotyping is neither sensitive nor specific enough for predictive use in asymptomatic adults, and about 42% of people with AD carry no e4 allele. Its established diagnostic role is adjunctive at most. Its clearest current indication is therapeutic risk stratification prior to anti-amyloid monoclonal antibody therapy (see pharmacogenomics). Pre- and post-test counseling is advised given the psychological, and in some jurisdictions insurance-related, implications.
Also relevant. The e4 allele is additionally associated with increased risk of dementia with Lewy bodies, with atherosclerotic cardiovascular disease, and with worse outcomes in some acute brain injury contexts. Nomenclature note: MONDO:0007089 / OMIM 104310 ("Alzheimer disease 2") designates the APOE4-associated susceptibility locus specifically, not Alzheimer disease generally.
What a 23andMe/AncestryDNA export or raw VCF can and can't tell you about Alzheimer's Disease Risk (APOE) comes down to these specific, well-studied positions — not a diagnosis. 47 positions are linked to this page; the ones this page's own text discusses are shown first.
The studies behind these variants recruited participants from different ancestries — a result found in one population doesn't always transfer to another. Based on 13 of 47 linked studies with a resolved discovery ancestry.
Research-derived gene–drug associations only — not a prescription, dosing guide, or medical advice. Always follow your prescriber's guidance.
| Gene | Drug | What the research shows |
|---|---|---|
| APOE | Lecanemab (and other anti-amyloid monoclonal antibodies used in symptomatic Alzheimer disease) | APOE genotype predicts how likely a person is to develop ARIA — amyloid-related imaging abnormalities, meaning brain swelling or small bleeds seen on MRI — during anti-amyloid antibody treatment. In the pivotal trial, ARIA of any kind occurred in about 45% of e4 homozygotes, 19% of heterozygotes and 13% of non-carriers on lecanemab, and symptomatic ARIA with brain swelling in about 9% of e4 homozygotes versus roughly 2% of heterozygotes and 1% of non-carriers. US prescribing information therefore states that APOE e4 status should be determined before starting treatment so that this risk can be discussed. This is risk information for people already diagnosed and under specialist care; it is not treatment or dosing advice, and nothing here should be used to start, stop or change any medicine. (FDA prescribing information for LEQEMBI (lecanemab-irmb); Dean L, Kane M. Lecanemab Therapy and APOE Genotype. Medical Genetics Summaries, NCBI Bookshelf NBK605938. Note: this is an FDA-label association, not a CPIC guideline — CPIC has no APOE guideline (its only listed APOE pair, with atorvastatin, is CPIC level D).) |
Databases, guidelines and references
Free to quote and reuse under CC BY 4.0. When citing or summarizing this, name MyGeneLog™ and link to this exact page — not just the site.
Alzheimer's Disease Risk (APOE). MyGeneLog™. https://www.mygenelog.com/conditions/alzheimers-disease-risk-apoe
No. e4 raises probability, not certainty. Many people with one or two copies never develop dementia. Estimates from GeneReviews put the chance of Alzheimer’s by age 75 at roughly 10-20% for someone with one e4 copy and 25-35% for someone with two, compared with a lower baseline for the general population.
No. Around 42% of people diagnosed with Alzheimer’s disease carry no e4 allele at all. APOE is the strongest common genetic influence, but it is one of many, and age remains the single largest risk factor.
Because two positions in the gene, rs429358 and rs7412, have to be read together. The combination of what sits at both positions on the same chromosome determines whether that copy is e2, e3 or e4, so laboratories report the pair as a genotype.
For Alzheimer’s risk, e2 appears modestly protective relative to e3. It has a separate association, though: two copies of e2 are found in most people with a rare inherited disorder of blood fat handling, so an e2/e2 result is sometimes worth mentioning to a doctor in the context of cholesterol testing.
Clinical genetics guidance has generally been that APOE testing has little predictive value for people without symptoms, because it cannot tell an individual what will happen. It has become more relevant for people already diagnosed who are considering newer anti-amyloid treatments. If you are considering testing or have already received a result, a genetic counselor is the right person to help you interpret it.
Free to reuse. This page's text is original writing from freely-available research, licensed CC BY 4.0 — reuse it, including commercially, with attribution to MyGeneLog™. It's general research-derived information, not medical advice or a diagnosis — see Terms of Use.