APOE ε4 and Alzheimer's risk, what your 23andMe result actually means

The APOE gene is the highest-effect common genetic variant we know of for late-onset Alzheimer's disease. If you've had your genome sequenced, through 23andMe, AncestryDNA, MyHeritage, or a clinical lab, your file contains your APOE status. About 25% of people of European ancestry carry one ε4 allele. About 2% carry two. APOE e4 Alzheimer's risk is the single most-searched topic in consumer genomics for good reason: the effect size is large enough that the result actually changes how you should think about the next 30 years.

If you're one of them, your relative risk of developing Alzheimer's is genuinely elevated compared to ε3/ε3 carriers. But "elevated relative risk" is a wildly imprecise framing. This post walks through what the actual numbers are, what the limits of that elevation are, and what, based on current evidence, carriers can actually do.

Everything below is anchored to the peer-reviewed literature. Nothing here is medical advice. If you carry ε4 and are worried about it, the right next step is a conversation with a physician, ideally one familiar with neurodegenerative-disease risk.

What APOE ε4 is, biologically

APOE encodes apolipoprotein E, a protein involved in cholesterol transport in the brain and elsewhere. The human gene comes in three common forms, ε2, ε3, and ε4, distinguished by two single-nucleotide differences. Two SNPs define which version you carry: the rs429358 APOE variant and the rs7412 APOE variant. The ε4 form differs from the reference ε3 by a single amino acid swap (cysteine to arginine at position 112), and that swap appears to change how the protein binds lipid particles. ClinVar's aggregated APOE submissions and the NHGRI-EBI GWAS Catalog APOE page are good starting points if you want the raw evidence rather than a summary.

In the brain, ε4 carriers show several measurable differences: reduced amyloid clearance, altered lipid metabolism in neurons, and modest changes in tau aggregation. These are mechanistic findings. They don't yet translate cleanly into "do X and your risk drops by Y."

What the numbers actually say

The risk numbers vary across studies based on age cohort, ancestry, and study design. The consensus picture from large meta-analyses (Genin et al. 2011; Reiman et al. 2020; and the IGAP consortium analyses, plus the broader PubMed literature on APOE e4 and Alzheimer's):

• ε3/ε4 heterozygotes: roughly 3x relative risk of late-onset Alzheimer's compared to ε3/ε3.
• ε4/ε4 homozygotes: roughly 8-15x relative risk, with the higher end of that range now believed to be a near-complete penetrance (most ε4/ε4 individuals will develop AD pathology if they live long enough).
• ε2 carriers: protective. ε2/ε2 is associated with reduced relative risk.

"Relative risk" sounds dramatic but the absolute numbers matter more. The lifetime risk of AD in the general population is roughly 10-15%. With one ε4, that becomes roughly 25-30% by age 85. With two ε4s, current best estimates put lifetime risk above 50%, with substantially earlier onset.

A few critical caveats:

• These numbers are population averages. They don't predict any individual's outcome. About a quarter of ε4/ε4 homozygotes never develop AD pathology. About a tenth of AD cases have no ε4 at all.
• Ancestry matters. Most numbers above derive from European-ancestry cohorts. The penetrance of ε4 appears lower in African-ancestry populations and somewhat different in East Asian populations.
• The risk is age-dependent. ε4 shifts the age of onset curve more than it shifts the binary "will develop AD" outcome. An ε4 carrier whose Alzheimer's would have onset at age 80 might see it at age 70 instead.

What carriers can actually do

This is where the literature gets messier and the temptation to recommend things on thin evidence gets strongest. The honest version:

Strongly evidence-anchored interventions:

• Cognitive screening starting earlier. Most major neurology bodies recommend baseline cognitive screening at 50-55 for ε4/ε4 homozygotes and around 60 for ε3/ε4 heterozygotes, vs. 65 for the general population. The intervention isn't curative; the rationale is earlier detection, earlier counseling, and earlier enrollment in clinical trials if desired.
• Cardiovascular risk control. Vascular disease and Alzheimer's pathology interact: vascular damage compounds amyloid pathology. The CAIDE risk score (Kivipelto et al. 2006 and updates) treats midlife hypertension, hypercholesterolemia, obesity, and diabetes as modifiable AD risk factors. The effect sizes for treating these are modest individually but meaningful when stacked.

Moderately evidence-anchored interventions:

• Regular aerobic and resistance exercise. The FINGER trial (Ngandu et al. 2015) and several follow-on studies show modest cognitive benefit from multidomain lifestyle intervention including exercise. Effect sizes are real but small. Not a cure; not nothing.
• Mediterranean / MIND diet. Observational evidence is consistent. RCT evidence is thinner but supportive. Less specific to ε4 than the general AD prevention literature.
• Sleep quality. Glymphatic clearance of amyloid happens preferentially during deep sleep. Poor sleep is plausibly causal for amyloid accumulation in ε4 carriers; treating sleep apnea and prioritizing sleep duration has plausible mechanism plus observational support.

Things the literature does NOT support, despite frequent claims:

• Specific supplements as primary prevention. Omega-3 trials in AD prevention have been disappointing. B-vitamin trials for homocysteine reduction haven't shown cognitive benefit. Curcumin, resveratrol, and similar antioxidant supplements have shown nothing meaningful in well-powered RCTs.
• Coconut oil, ketogenic diets, and metabolic-flexibility interventions. Promising preclinical data, no convincing clinical-trial evidence.
• "APOE-personalized" anything. The supplement industry has rushed to claim that specific compounds work better or worse for ε4 carriers. The actual evidence base for differential efficacy is thin to nonexistent.

The right framing is: the interventions that work for general dementia prevention are the same interventions that work for ε4 carriers. ε4 status changes the urgency of doing them, not the what.

What to do if you just learned you're ε4-positive

A few concrete next steps that almost all neurology consensus statements endorse:

1. Don't panic. The increased risk is real but population-level, and it's distributed across decades of life. Plenty of ε4 carriers live to advanced age cognitively intact.
2. Talk to a primary-care doctor. Mention your ε4 status. Ask about cardiovascular risk screening (BP, lipid panel, A1C) and baseline cognitive assessment timing.
3. Sort out cardiovascular risk first. If your blood pressure or lipids are out of range, treating those is the highest-leverage thing you can do right now.
4. Lifestyle basics. Exercise (aerobic + resistance), sleep, Mediterranean-pattern eating. None of these are silver bullets. All of them are evidenced to delay cognitive decline modestly.
5. Consider counseling. Genetic counselors are trained for exactly this situation. If you have family members who might also carry ε4, a counselor can help frame the conversation.

What's coming next

The AD clinical-trial pipeline has had some of its most active years recently. Lecanemab and donanemab (anti-amyloid antibodies) have been approved with meaningful but modest cognitive-decline-slowing effects. The Anti-amyloid Treatment in Asymptomatic AD (A4) trial is testing earlier intervention. The TANGO and LATAM studies are looking at ε4-stratified responses to these therapies. The picture in five years will look different from the picture today.

For carriers, the practical implication is: the interventions of 2030 may be substantially better than the interventions of 2025. That makes early monitoring valuable not just for prevention but for trial eligibility. The framing we like for thinking about APOE and longevity is that ε4 status changes the slope of the cognitive-aging curve, not its terminus, and the levers that bend the slope are mostly available today.

A note on Expressive

If you have your 23andMe / AncestryDNA / MyHeritage raw file and want a citation-anchored interpretation of your APOE status (plus the other 600,000+ variants in the file), you can sign up here. Every claim links to the underlying study; we tell you when the evidence is thin; we don't recommend supplements that the literature doesn't support. APOE is one of the variants we have particularly thorough citation coverage on, the rs429358 variant page is a good place to look at how we surface the underlying evidence.

If you're already on Expressive and you carry ε4, the variant page surfaces the same kind of layered evidence summary you'd find here, plus the integration with your other carried variants, because ε4 risk interacts with cardiovascular variants in ways worth understanding together. We don't prescribe, we describe: every number on the page links to the study it came from, and your genome stays yours.

Want updates when we ship new variant pages or a research deep-dive? Read the latest issue or get notified about early access.