ALDH2 Deficiency (Alcohol Flush Reaction)

Screening · §95

If you flush red after one beer and your family is from East Asia, you almost certainly carry a broken copy of a gene called ALDH2. The flush is your body telling you the alcohol you just drank is sitting inside you half-processed — as a chemical called acetaldehyde — and acetaldehyde is a known carcinogen. Carriers who drink regularly run several times the normal risk of esophageal and throat cancers.

Test · Once Evidence Strong Chapter Screening

Roughly 540 million people carry this — about one in three Han Chinese, Japanese, and Korean adults (Chen et al. 2014). The big payoff to knowing is on the cancer side: a heterozygous carrier who drinks two beers a day from his twenties through his fifties carries esophageal-cancer odds comparable to a heavy-drinking non-carrier. The screen is cheap — two questions you can answer right now, or a hundred-dollar genetic test if you want to be sure. The catch is the action it asks of you: if you carry it, the safest drinking dose is zero.

Drinking alcohol is a two-step chemistry problem inside your body. Step one: an enzyme in your liver turns the ethanol from your drink into a chemical called acetaldehyde — toxic, reactive, and unpleasant to be near. Step two: a second enzyme, ALDH2, turns the acetaldehyde into acetate, which is harmless. Acetate is what your body wanted all along.

If you carry the broken version of ALDH2 — the one with a single typo in the gene called rs671 — step two breaks. The acetaldehyde piles up in your blood and stays there for hours instead of clearing in minutes. That's the flush: your face going red, your heart racing, the dull headache, the slight nausea. It's your body responding to a small chemical fire it can't put out.

About four in five East Asians also carry a second variant — in the first enzyme, the one that makes the acetaldehyde — that produces the toxic intermediate faster than normal (Chang et al. 2017). Both genes leaning the same direction is what makes the East Asian flush so much sharper than the occasional European who turns pink on red wine.

The cancer risk isn't a guess

Acetaldehyde is on the same World Health Organization list as tobacco smoke and asbestos — a category-one human carcinogen (IARC 2012). Carriers who drink keep it in contact with the lining of their mouth, throat, and esophagus for hours at a time, and the chemistry damages DNA in those tissues directly.

The number that should sit with you: a heterozygous carrier who drinks at the same level as a non-carrier carries roughly three times the odds of esophageal cancer.

One counterintuitive twist: people with two broken copies — the homozygotes — actually have lower observed cancer rates than heterozygotes (Lewis & Smith 2005). Not because the chemistry is kinder. Because the reaction is so violent — sweating, vomiting, racing heart from a single sip — that almost no homozygote ever becomes a regular drinker. The cancer burden lands on the people in the middle: the ones with one broken copy, who can push through the flush and keep drinking for decades.

What sustained drinking looks like, in the carrier's body

For the heterozygous reader who has been "the friend who turns red" through college, business school, and a corporate career — drinking moderately, two or three beers a night a few nights a week, with the occasional client dinner — the next few decades aren't dramatic until they are. The throat clearing in your forties that you put down to acid reflux. The hoarse voice for a week after a heavy weekend that you assume is a cold. The first endoscopy in your fifties that finds something the gastroenterologist needs to talk about in person.

Esophageal squamous-cell cancer is the one to be most afraid of. Caught late — which is how it's usually caught — the five-year survival is grim, and the carrier-drinker statistical risk profile is closer to that of a heavy drinker than to that of your non-carrier coworker pouring the same amount (Yokoyama et al. 2003). The cancers of the larynx and throat that travel with it carry similar lethality (Boccia et al. 2009).

Short of cancer, the lower-grade story is that chronic carrier-drinkers run higher rates of macrocytic anaemia — red blood cells that come out the wrong size because the acetaldehyde is interfering with how they're built (Yokoyama et al. 2019). You feel tired in a way that doesn't track to sleep.

How to find out, in two questions

The cheapest screen in medicine, almost. Two questions, validated to be roughly as good as a genetic test:

If your drinking history is ambiguous — you didn't drink in your twenties because the reaction was unpleasant, you don't remember clearly, or you're just curious — a single-site genetic test for the rs671 variant settles it. 23andMe, AncestryHealth, and the cheaper direct-to-consumer kits all include it. Cost is roughly a hundred to two hundred dollars once, and the result is essentially yes-or-no.

If the answer is yes — what to do with it. The honest answer is that the safest drinking amount for a carrier is zero. There is no pill, no charcoal tablet, no glutathione gummy that lets you metabolize acetaldehyde faster; the enzyme is broken, and nothing on the shelves fixes it. The operational compromise most carriers land on is "rare ceremonial drink" — a wedding toast, a New Year's glass — and otherwise none. The further you are from that and the closer to weekly drinking, the bigger the lifetime cancer accumulator runs.

What most people get wrong about the flush

"It's a cosmetic problem." The redness on your face is the visible piece of a chemical event that's happening throughout your bloodstream and along the entire lining of your throat. The face is just where it shows. Suppressing the visible redness doesn't suppress what's happening to the tissues you can't see.

"I built up a tolerance." A lot of long-term carrier-drinkers report that the flush dampens after years of regular drinking. The visible reaction shrinks; the acetaldehyde clearance does not. What's actually happening is that the downstream histamine response is desensitizing — your blood vessels stop dilating as obviously — while the enzyme is still broken and the carcinogen is still pooling for hours (Brooks et al. 2009). The faded flush is the alarm getting quieter, not the fire going out.

"I'm not East Asian, so this doesn't apply." The classic rs671 variant is concentrated in East Asia, but newer work has found related variants — different single-letter typos, same broken-enzyme outcome — at meaningful rates in Latin American, South Asian, and African populations (Chen et al. 2014). Anyone who consistently goes pink from one drink should treat the question as live for them.

"Pre-drink supplements help." The activated charcoal, B-vitamin combinations, and "Asian flush relief" pills sold at convenience stores in Tokyo, Seoul, and Taipei have weak or no evidence for reducing acetaldehyde exposure (USC News 2020). Most work by blunting the visible flush — same problem as famotidine.

Who specifically should be running this check

Highest priority, ranked by how likely you are to carry it:

You're of East Asian heritage — Han Chinese, Japanese, Korean, Taiwanese, Vietnamese. Carrier rate is somewhere between one in three and one in two (Chen et al. 2014).
You go red after a single drink — any heritage. That symptom alone makes the variant likely enough to treat as a working assumption.
One of your parents flushes — fifty-fifty inheritance per copy. Worth asking at the next family dinner.
You're an East Asian adult who doesn't drink — and you don't know whether it's because you never liked the taste or because your first college beer made you violently sick. The screen is still worth running, both for the nitroglycerin caveat below and because a positive result is something to tell your kids.

If you've already stopped drinking, the screen still has two uses. First, the variant changes how some heart medications work — specifically, sublingual nitroglycerin for chest pain doesn't work as well in carriers because the same enzyme is responsible for activating it (Gross et al. 2015). Tell a future cardiologist. Second, the result is heritable information your children may want.

Where this goes wrong in practice

Reading the flush as embarrassment, not chemistry. The single biggest failure is the teenager and twenty-something who learns to push through the visible reaction so they can drink with their friends. The flush is doing exactly what it's evolved to do — telling them stop — and they spend a decade overriding the signal.

Quitting in your twenties, restarting in your thirties. Many carriers stop drinking in their teens or twenties because the reaction is unpleasant, then resume in their thirties or forties for professional reasons — client dinners, business trips, social mobility. The cancer-risk clock runs on cumulative lifetime exposure, so a late-start drinking career still accumulates years, just on a slower curve.

The "I don't flush anymore" trap. Long-term carrier-drinkers often report the visible reaction has faded. As above, the chemistry hasn't — the alarm has gone quiet, not the fire — and the absence of visible flush in someone with a known carrier history is not reassurance.

Masking with antihistamines. The H2-blocker workaround (famotidine, ranitidine) has gone mainstream in some Asian-American drinking scenes (USC News 2020). It produces the worst possible combination: heavier drinking without the deterrent symptom.

Medication notes for carriers

Two prescription drugs interact with the same broken enzyme. Worth knowing about even if neither is on your radar today — both matter on the day they suddenly are.

What changes when a carrier stops drinking

The first thing that changes is the recovery curve from a single social event. The Sunday morning after a Saturday wedding stops involving the dull headache and the racing-heart half-sleep you used to read as "I just don't tolerate alcohol well." Without the acetaldehyde load, hangover severity drops faster and further in carriers than in non-carriers, because there was more to drop.

The medium-term shift is social. The carrier who declines a drink at a work dinner with "I'm one of those flushers, I get cancer risk from it" gets a different reaction than the one who declines with "I'm not really drinking these days." The first is a medical fact; the second sounds like a phase. In East Asian professional cultures where drinking is woven into the workday, the medical framing buys real social permission that the lifestyle framing doesn't.

The long-term payoff is the cancer risk that doesn't materialize. The five-year esophageal-cancer survival statistic you read in the stakes section above stops being relevant to your decade-by-decade probabilities — gradually, because cumulative exposure matters and the curve doesn't fully reset for the first decade off alcohol, but meaningfully. At population level, Brooks et al. argue that universal screening of East Asian-heritage drinkers plus serious behavioural adoption would prevent a substantial fraction of esophageal cancers in this group — comparable in scale to a major tobacco-cessation campaign (Brooks et al. 2009).

Adjacent topics worth a look: general alcohol risk thresholds (the conversation about whether any amount of drinking is "safe" is its own argument and applies to non-carriers too); upper-GI endoscopic surveillance once you've been a heavy drinker for years; and the broader pharmacogenetics conversation about how ancestry-specific variants affect everything from beta-blocker response to anaesthesia. The development of drugs that can re-activate the broken enzyme (the Mochly-Rosen lab's Alda compounds) is real science but not yet a consumer story.

Related in the handbook

— This flush gene is exactly why alcohol's carcinogen risk hits some people much harder.
— A genetic test confirms the flush gene if the two-question check leaves you unsure.
— Both stack the same esophageal-cancer risk — if you carry the flush gene, also skip scalding drinks.

Substance and claimed effects

ALDH2 deficiency is the phenotype of carrying one or two copies of the rs671 A allele (Glu504Lys, historically Glu487Lys; the ALDH2*2 designation) in the mitochondrial aldehyde dehydrogenase 2 gene on chromosome 12. ALDH2 is the principal enzyme oxidising acetaldehyde — the first-pass metabolite of ethanol — to acetate. Because the functional enzyme is a tetramer, even one variant subunit poisons the complex: heterozygotes (ALDH2*1/*2) retain only under 20% of wild-type activity, and homozygotes (ALDH2*2/*2) are essentially null (Brooks et al., PLoS Medicine 2009)(Chang et al., J Biomed Sci 2017). The variant is concentrated in East Asia (Han Chinese, Japanese, Korean) at allele frequencies producing 30–50% carrier prevalence, with ~540 million carriers worldwide — roughly 8% of the global population — and a sharp west-to-east gradient that drops to ~0% in Europeans and most Africans (Li et al. 2009)(Chen et al., Physiol Rev 2014). The directly claimed phenotype is the alcohol flushing reaction — facial erythema, tachycardia, headache, nausea — driven by acetaldehyde accumulation. The downstream consequences in scope for this entry are: severity of hangover; markedly elevated risk of upper-aerodigestive-tract squamous cell cancers (esophageal, oropharyngeal, laryngeal, hypopharyngeal) when carriers drink; impaired bioactivation of nitroglycerin; smaller, less-settled signals on osteoporosis, Alzheimer's-disease risk, and coronary disease; and the screening implication that a simple two-item flushing questionnaire or a single-site genotype identifies carriers cheaply and lets them make an informed alcohol decision.

Evidence by addressing question

mechanism

Ethanol is oxidised by alcohol dehydrogenase (chiefly ADH1B in the liver) to acetaldehyde, then by ALDH2 to acetate. Acetaldehyde is genotoxic — it forms DNA adducts (N²-ethylidene-2'-deoxyguanosine, propanodeoxyguanosine), induces DNA double-strand breaks and sister-chromatid exchanges, and is classified by IARC as a Group 1 carcinogen "associated with the consumption of alcoholic beverages" (IARC Monograph 100E, 2012). In ALDH2*2 carriers the acetaldehyde clearance half-life is multiples longer; salivary acetaldehyde — the tissue dose most relevant to esophageal exposure — runs 2–3× higher in carriers after a fixed alcohol dose (Brooks et al. 2009). The Glu504Lys substitution disrupts the active-site geometry; because ALDH2 is a homotetramer, one mutant subunit blocks function across the whole assembly (dominant-negative), which is why heterozygotes carry far more than 50% of the homozygous penalty (Chen et al. 2014). Roughly 80% of East Asians also carry the high-activity ADH1B*2 allele (rs1229984), which accelerates ethanol → acetaldehyde, compounding the bottleneck (Chang et al. 2017). Beyond ethanol, ALDH2 metabolises a broader class of reactive aldehydes — including 4-hydroxynonenal from lipid peroxidation and nitroglycerin's bioactivation step releasing nitric oxide — which is the proposed mechanistic bridge to the non-alcohol disease associations (Chen et al. 2014)(Chen et al., Science 2008).

evidence

Esophageal squamous cell carcinoma (ESCC). The Lewis & Davey Smith meta-analysis pooling Mendelian-randomization-style case-control data found heterozygous (*1/*2) drinkers had an odds ratio of 3.19 (95% CI 1.86–5.47) versus *1/*1 drinkers at equivalent alcohol intake; homozygotes (*2/*2) had OR 0.36, but only because they almost never drink at all (Lewis & Smith, Cancer Epidemiol Biomarkers Prev 2005). Yokoyama's Japanese-male cohort attributed 68.5% of ESCC excess risk in inactive-heterozygous drinkers specifically to drinking, with population-attributable fractions among the highest reported for any gene-environment interaction in cancer (Yokoyama et al. 2003). A 2023 meta-analysis confirmed the heterozygote-drinker risk magnitude and showed the signal is robust to publication-bias correction (Zhang et al., Cancer Med 2023).

Head and neck squamous cell carcinoma. The Boccia Mendelian-randomization meta-analysis found ALDH2*2 heterozygote drinkers carry markedly elevated risk for oropharyngeal, laryngeal, and hypopharyngeal cancer, with effect sizes paralleling the esophageal signal (Boccia et al. 2009). Yokoyama's series of Japanese alcoholic men documented sharply elevated rates of multiple primary cancers across the upper aerodigestive tract among ALDH2*1/*2 carriers (Yokoyama et al. 2003).

Flushing severity and hangover. Blood acetaldehyde concentration at 4 hours after a standardised dose correlates strongly with hangover severity; ALDH2*2 carriers show 6× higher peak acetaldehyde than non-carriers and report dose-for-dose worse hangovers (Brooks et al. 2009). The flushing reaction itself involves acetaldehyde-triggered histamine release, vasodilation, and catecholamine surge — hence the tachycardia and headache.

Nitroglycerin response. Nitroglycerin's anti-anginal effect depends on ALDH2-mediated denitration to release nitric oxide; ALDH2*2 carriers show attenuated coronary vasodilation and faster tolerance development, with clinical implications when these patients present with angina (Chen et al., Science 2008)(Gross et al. 2015).

Other associations (weaker, partly observational). Meta-analyses signal modest associations of ALDH2*2 with coronary artery disease in East Asian populations, with Alzheimer's-disease risk, and with osteoporosis / hip fracture — proposed mechanism being unchecked endogenous aldehyde load (4-HNE, lipid-peroxidation products). Evidence is consistent in direction but smaller in magnitude than the alcohol-cancer signal and partly confounded by lifetime drinking history (Chen et al. 2014).

protocol

The screening protocol is two-step. Step 1, identification. The Yokoyama two-item flushing questionnaire — "Do you flush in the face immediately after a single glass (180 mL) of beer?" and "Did you flush during your first year or two of drinking?" — is roughly 90% sensitive and 90% specific for ALDH2 deficiency in Japanese adults (Yokoyama et al. 2003). A single-site genotype (rs671) — available on 23andMe, Ancestry, and direct-to-consumer genetic kits — is essentially deterministic and the gold standard when the flushing history is ambiguous (about a third of East Asian Americans report lifetime abstention, so they have no flushing history to query). Step 2, action. Carriers should sharply limit or eliminate alcohol. The Brooks et al. PLoS Medicine recommendation is explicit: clinicians should counsel ALDH2-deficient patients that even moderate drinking carries materially elevated cancer risk (Brooks et al. 2009). Concretely, a heterozygote drinking ~2 standard drinks/day has esophageal-cancer odds comparable to a heavy-drinking non-carrier; the safest dose for a carrier is zero, with "occasional ceremonial drink" the operational compromise.

contraindications

The variant itself is the contraindication for regular alcohol. Additionally: ALDH2*2 carriers presenting with chest pain should not be assumed to respond to sublingual nitroglycerin at standard doses — the bioactivation is impaired and tolerance develops faster (Gross et al. 2015). Disulfiram, which inhibits ALDH2 to enforce alcohol aversion, is functionally redundant in *2/*2 carriers and can produce a severe reaction in heterozygotes who drink.

misconceptions

The dominant misconception is that flushing is a cosmetic embarrassment to be hidden. The flushing is the body's chemical alarm: it is the visible expression of acetaldehyde concentrations that are doing covert damage to esophageal epithelium. Suppressing the flush with an H2 blocker (famotidine, ranitidine) — a widespread folk remedy — addresses only the histamine-driven redness while leaving acetaldehyde clearance unchanged, allowing the carrier to drink more without the warning signal (USC News 2020). The University of Southern California Keck School public-health communication is explicit that this is a dangerous practice; it produces the maximally bad combination of elevated alcohol intake without the deterrent symptom (USC News 2020).

A second misconception: "I'm not East Asian, so this doesn't apply." About 8% of the global population carries rs671 or one of the newly characterized variants (P92T, V304M in Latin Americans; T224M in South Asians; I41V in some African populations) that produce similar acetaldehyde clearance deficits (Chen et al. 2014). Anyone who flushes consistently from one drink should treat the screening question as live.

A third: "Homozygotes are at the highest cancer risk." Counterintuitively, *2/*2 homozygotes have the lowest observed esophageal cancer rates because the reaction is severe enough to enforce near-abstinence. The highest population cancer burden falls on heterozygotes, who can drink heavily enough to accumulate years of acetaldehyde exposure (Lewis & Smith 2005)(Yokoyama et al. 2003).

audience

Highest-yield screening targets, by descending pre-test probability: East Asian ancestry (~30–50% carrier rate); anyone with a personal history of flushing after one drink (regardless of ancestry); first-degree relatives of confirmed carriers (50% transmission per copy). Older male East Asian drinkers carry the largest absolute cancer burden because the carcinogenic effect is cumulative and most observational data is in men (Yokoyama et al. 2003). For non-drinkers the screening result has limited downstream consequence except for the nitroglycerin caveat and the relevant family history.

alternatives

There is no pharmacologic rescue for the carrier who wants to drink "safely." Activator molecules (Alda-1 and successors developed by the Mochly-Rosen lab at Stanford) restore ALDH2 activity in pre-clinical models and have proof-of-concept human data, but they are not available as consumer products and would in any case treat the symptom, not eliminate it (Chen et al. 2008)(Chen et al. 2014). Antihistamines, charcoal, glutathione precursors, and B-vitamin combinations sold as "Asian flush" supplements have weak or no evidence for reducing acetaldehyde exposure and several are explicitly dangerous because they obscure the flush deterrent.

failure-modes

The dominant practical failure is selective denial — flushing is read as a low-stakes nuisance rather than a carcinogenic exposure marker, and the carrier continues drinking at "social" levels for decades. A second failure: at-risk individuals who don't drink during their flushing-prone late teens / early twenties (because the reaction is uncomfortable) but resume in their thirties / forties for professional or social reasons, accumulating cancer risk on a delayed exposure curve. A third: heterozygotes who develop apparent tolerance — the visible flush dampens over years of chronic drinking through downregulation of mediators — and read the dampened reaction as a clean bill of health, when acetaldehyde exposure is unchanged (Brooks et al. 2009).

practicalities

The two-item flushing self-screen is free and takes thirty seconds. Direct-to-consumer genetic testing including rs671 costs roughly $100–200 once and is available from 23andMe, AncestryHealth, and similar platforms. Clinical genotyping through a primary-care order is similar in cost and increasingly covered when there is an East Asian-heritage indication. The behavioural cost of "if positive, don't drink" is real but modest in cultures and contexts where teetotalism is available; it can be substantial in social, professional, or family contexts where drinking is normative (Japanese / Korean / Chinese business culture, family celebrations).

history

The flushing reaction has been recognised clinically since at least the 1970s in East Asian populations; the underlying ALDH2 polymorphism was characterised molecularly in the 1980s. The geographic-distribution work by Li et al. tracked the variant to a likely Han Chinese origin roughly 2,000–3,000 years ago, with subsequent expansion via population movements; one hypothesis links its persistence to a selective advantage against pathogens in rice cultivation, though selection-pressure evidence is debated (Li et al. 2009). The contemporary public-health framing — flushing as cancer-risk marker rather than nuisance — was crystallised by the Brooks et al. PLoS Medicine review in 2009 and the parallel NIAAA statement (Brooks et al. 2009)(NIAAA 2009).

stakes

For a heterozygous East Asian male drinking moderately from his 20s through his 50s, the absolute lifetime risk of esophageal squamous cell carcinoma is meaningfully elevated — Yokoyama's population-attributable-fraction work suggests two-thirds of ESCC in this group is alcohol-attributable, with five-year survival of diagnosed ESCC around 20% (Yokoyama et al. 2003). Head-and-neck cancers (laryngeal, oropharyngeal) carry similar lethality and parallel ALDH2*2-drinking risk (Boccia et al. 2009). Even short of cancer, sustained drinking by carriers drives macrocytic anaemia and chronic acetaldehyde exposure to oral and esophageal mucosa (Yokoyama et al. 2019).

payoff

For a confirmed carrier who stops drinking, the elevated cancer risk attenuates over time but does not fully reset within the first decade; cumulative exposure matters. The immediate payoffs are felt sooner: no more flushing or palpitation episodes, materially reduced hangover severity from any inadvertent exposure, and — at the population level — the structural removal of a major modifiable cancer risk factor. Brooks et al. estimate that universal screening of East Asian-heritage drinkers plus behavioural adoption could prevent a substantial fraction of esophageal cancers in this population (Brooks et al. 2009).

out-of-scope

Related but outside this entry: general alcohol minimization for non-carriers (its own entry on alcohol risk thresholds); upper-GI endoscopic screening protocols for established heavy drinkers (esophageal cancer surveillance); the broader question of ADH1B*2 status and its independent contribution; the developing pharmacology of ALDH2 activators (research-grade, not consumer-actionable); fetal alcohol exposure in maternal carriers; alcohol-use-disorder treatment.

The credibility range

Optimist case

The mechanistic, epidemiological, and clinical evidence converge with unusual cleanliness: a single SNP produces a measurable biochemical defect, the biochemical defect produces a quantifiable cancer-risk elevation in carriers who drink, the cancer-risk elevation is replicated across multiple Asian populations and confirmed by Mendelian-randomization designs that bypass the usual observational confounders, and the screening intervention (a two-question survey) has high sensitivity and specificity. This is one of the cleanest gene-environment interactions in oncology. The IARC Group 1 classification of acetaldehyde in alcoholic beverages is reinforcing rather than load-bearing — the data would carry the recommendation independently. For East Asian populations this is a major preventable cancer burden, comparable in scale to tobacco-attributable lung cancer; universal screening plus carrier abstinence would shift national esophageal-cancer incidence curves measurably (Chang et al. 2017).

Skeptic case

The Lewis & Smith Mendelian-randomization meta-analysis flagged that the *1/*2 vs *1/*1 cancer-risk odds ratio is partly confounded by the *1/*2 carriers' altered drinking behaviour — they don't drink the same way as *1/*1 controls at the same reported quantity, complicating the gene-environment interaction picture (Lewis & Smith 2005). The 2023 Zhang Mendelian-randomization analysis found a strong observational signal but a null or attenuated causal estimate using genetic instruments, suggesting the magnitude in published case-control series may be inflated (Zhang et al. 2023). Non-cancer associations (Alzheimer's, osteoporosis, CAD) are smaller in magnitude, partially confounded by lifetime alcohol exposure, and not yet ready for clinical action. The Yokoyama Japanese cohorts are heavily male and heavily alcoholic; effect-size generalisation to female, light-drinking, or non-Japanese East Asian populations is empirically thinner. Western clinical genetics has not adopted the screening practice at scale, perhaps reflecting genuine epidemiological caution rather than mere inertia.

Author's call

The esophageal- and head-and-neck-cancer signal is solid: meta-analyses converge, the mechanism is biochemically explicit, the Mendelian-randomization caveats narrow the effect size somewhat but do not erase it, and the absolute risk in a heterozygous East Asian male drinking from his 20s is materially elevated regardless of which estimate one prefers. The non-cancer associations are real-but-tentative and should not drive the recommendation. The recommendation that crystallises: an East Asian-ancestry reader who flushes from one drink should treat the variant as a working diagnosis, confirm cheaply (questionnaire and/or rs671 genotype), and reduce alcohol intake materially — ideally to zero, operationally to "rare ceremonial." This is a high-evidence, low-controversy recommendation; evidence reads 4–5 and controversy reads 1.

Stakeholder and incentive map

Patient-advocacy / research groups (Stanford's ALDH2 group, the Mochly-Rosen lab) — push the screening framing and the precision-medicine angle. Incentive: scientific and humanitarian.
Public-health bodies in East Asia (Japanese cancer-prevention efforts, Taiwan / Korea cancer registries) — increasingly recognise ALDH2 status as a primary-prevention lever; uptake is uneven and runs against deep alcohol-culture norms (Chang et al. 2017).
Alcohol industry — has not engaged substantively; the population-attributable burden is concentrated in a minority of carriers, limiting commercial salience.
Consumer-supplement industry — markets "Asian flush relief" products (charcoal, B-vitamins, histamine blockers, glutathione precursors), most of which have no acetaldehyde-clearance evidence and several of which (H2 blockers especially) are actively harmful to risk because they mask the deterrent signal (USC News 2020).
Direct-to-consumer genetics (23andMe, Ancestry) — include rs671 in standard panels. Incentive: aligned with carrier-identification but not driving clinical uptake.
Western clinical genetics — has been slow to adopt ALDH2 screening despite high pre-test probability in Asian-heritage patients; reflects general under-engagement with population-specific pharmacogenetics.

Population variability

Ancestry. Carrier frequency ~30–50% in Han Chinese, Japanese, Korean; 10–18% in Vietnamese; near-zero in Europeans, sub-Saharan Africans, most South Asians. Non-rs671 ALDH2 variants with similar acetaldehyde-clearance phenotypes have been reported in Latin American, South Asian, and African populations but are individually rarer (Li et al. 2009)(Chen et al. 2014).
Sex. Most large cohort data is male; women with the variant show the same biochemical phenotype but lower observed absolute cancer rates, partly explained by lower historical drinking exposure. Mechanism-based risk does not change.
Age. Risk is cumulative; the variant matters most for those still in their drinking years. Once exposure stops, risk plateaus rather than fully reverting.
Co-variant ADH1B*2. ~80% of East Asians carry the rapid-converter ADH1B*2 allele, which compounds the acetaldehyde bottleneck; effect on cancer risk in combination with ALDH2*2 is multiplicative (Chang et al. 2017).
Tobacco co-exposure. Smoking with ALDH2*2 substantially compounds upper-aerodigestive cancer risk; many of the highest-OR Yokoyama series cells are smokers (Yokoyama et al. 2003).

Knowledge gaps

Outstanding questions: the magnitude of non-cancer disease associations (Alzheimer's, CAD, osteoporosis) once lifetime alcohol exposure is fully accounted for; the safe drinking threshold (if any) for a young heterozygote — current best-evidence default is "zero," but absolute risk per drink-year is imprecisely characterised; whether ALDH2 activators in development will alter the screening calculus by offering a pharmacologic rescue route; the absolute population impact of universal screening at population scale in Japan or Korea (modelling exists; outcome data does not); the safe alternative for ALDH2-deficient angina patients who would normally receive sublingual nitroglycerin (current clinical default is to use the drug but expect attenuated response, with no large outcome trials guiding alternatives); whether the dampening of visible flush over chronic exposure reflects any true protective adaptation or purely a histamine-axis change without acetaldehyde clearance.

Scope vs. brief. The brief named flushing, hangover severity, esophageal and head-and-neck cancer risk, and screening implications. All four are covered end-to-end. Two additional consequences fell out of the dossier and got named in the article rather than dropped: the impaired nitroglycerin response (mechanism-relevant and a real clinical caveat for older carriers) and the macrocytic-anaemia signal in chronic carrier-drinkers. The smaller, weaker signals on Alzheimer's disease, coronary artery disease, and osteoporosis were excluded from reader prose by design — magnitudes are modest, alcohol-exposure confounding is hard to net out, and clinical action does not yet differ from the cancer-driven recommendation. They are noted in the research dossier's evidence section so a future reviewer can see where the boundary was drawn.

Hard scoping calls.

Action: test rather than know or avoid. The reader's prescribed move is to determine their genotype (two questions, or a $100 kit) and act on a positive. avoid would only fit the second half; know understates the active screening recommendation. test captures the gather-your-own-data character that drives this entry.
Cadence: once. The variant doesn't change; one screen and the subsequent behavioural commitment is for life.
Category: screening over medical. The entry's load-bearing reader action is the cheap two-question identification, which is screening behaviour. medical would imply ongoing clinical management; this is a one-time literacy event with lifestyle downstream.
Audience scoping: left unset rather than narrowed to East Asian heritage. The variant is concentrated there but related broken-enzyme variants exist in Latin American, South Asian, and African populations, and the practical filter — "do you flush from one drink?" — is universal. Narrowing the audience field would mis-signal that non-Asian flushers should ignore the entry.

Rating difficulties.

longevity at 4 vs 5. The cancer-prevention signal is genuinely large for the heterozygous-East-Asian-drinker subpopulation, but the entry's longevity score has to read as the substance's effect across all readers — most of whom are non-carriers for whom the entry has zero longevity impact. 4 reflects "major effect for the audience this matters for" without overclaiming a 5 (dominant, population-level mortality shift).
effort_burden at 2. Knowing your status is trivial; sustaining abstention is real work, particularly in East Asian business-drinking culture. 2 splits the difference. A case for 3 exists if the reader is starting from heavy drinking, but the average-reader anchor in the framework argues against it.
mood at 1. The honest case is that confirming a long-running embarrassment as a medical fact (and getting a socially defensible reason to decline drinks) lifts mood marginally. Could defensibly be 0; 1 was the call because the "I have a name for this" effect is real and reported.

Future-link candidates. Once they exist: an entry on general alcohol risk thresholds for non-carriers; an entry on upper-GI endoscopic surveillance for chronic heavy drinkers; an entry on sublingual nitroglycerin and the broader pharmacogenetics of cardiovascular drugs in East Asian populations. The Alda-class ALDH2 activators are research-stage and don't warrant an entry yet.

Separate-entry candidate. ADH1B*2 is mentioned in passing in the mechanism section because it compounds the ALDH2*2 cancer-risk signal. It probably warrants its own short entry covering its independent effect on alcohol-use-disorder susceptibility, given that ~80% of East Asians carry it.