Skip to main content

Longevity Library · Lab reference

Written by Daniel Tagge, MD
Updated October 23, 2026
Reviewed October 24, 2026
51 sources · Measured in standard drinks per week

System: Exposomics

AlcoholEthanol · Drinking · Standard drinks

Across 83 prospective studies of 599,912 current drinkers, all-cause mortality was lowest at or below about 100 grams of alcohol a week, roughly seven US drinks, and life expectancy at 40 fell by about six months above that, one to two years above 200 grams and four to five years above 350 grams; the Global Burden of Disease analysis put the intake that minimized overall health loss at zero, and the World Health Organization states that no safe amount can be established. The apparent protection of moderate drinking against heart disease and stroke in the older cohorts largely disappears when drinking is instrumented by genes, when former drinkers are separated from lifetime abstainers, or when the drinkers' healthier habits are accounted for; the US guidelines ask adults who drink to keep to 2 drinks or less in a day for men and 1 or less for women, and do not recommend that anyone start.

Alcohol is the one common intake that moves half the markers on this shelf at once: GGT, ALT, ferritin, triglycerides, uric acid, HDL and blood pressure. The old J-curve, in which a drink or two a day looked protective, rests on cohort comparisons that the genetic studies and the bias-corrected meta-analyses do not confirm; the risks of cancer, atrial fibrillation, high blood pressure and stroke rise from low intakes without a threshold. Pattern matters as much as total, the markers answer within weeks, and a month off is the cleanest test of what drinking is doing to a particular person's numbers.

01 · The scale

Where the laboratory range and the evidence sit.

Lowest all-cause mortality at or below about 100 grams a week (about 7 US drinks) across 599,912 drinkers, with no threshold below which stroke, heart failure or hypertensive death stopped falling; the Global Burden of Disease minimum was zero. The J-curve for the heart does not survive Mendelian randomization in 512,715 Chinese adults or 371,463 Britons, and the mortality benefit of light drinking vanishes once former drinkers and study quality are accounted for. Cancer, atrial fibrillation and blood pressure rise with dose from low intakes; stopping lowers blood pressure and halves atrial fibrillation recurrence in a randomized trial.

07142128

  1. US guideline cap, drawn as a weekly sum: 1 or fewer a day for women (7 a week), 2 or fewer for men (14), on days alcohol is consumed

    0 to 14 standard drinks per week · US Dietary Guidelines 2020 to 2025

  2. Intake that minimized overall health loss: zero drinks a week

    0 to 0 standard drinks per week · GBD 2016 Alcohol Collaborators 2018

  3. Low risk: 2 or fewer standard drinks a week (13.45 g each)

    0 to 2 standard drinks per week · Canada's Guidance on Alcohol and Health 2023

  4. Moderate risk: 3 to 6 a week; cancer risk rising

    3 to 6 standard drinks per week · Canada 2023; Chen 2011; WHO 2023

  5. Lowest all-cause mortality at or below about 100 g a week, about 7 US drinks; stroke and blood pressure already rising

    0 to 7 standard drinks per week · Wood 2018; Millwood 2019

  6. Increasingly high risk from 7 a week; above 100 g a week about 6 months less life at 40, above 200 g 1 to 2 years

    7 to 14 standard drinks per week · Canada 2023; Wood 2018

  7. Heavy drinking by the NIAAA definition: 8 or more a week for women, 15 or more for men; above 350 g a week 4 to 5 years less life at 40

    15 and above standard drinks per week · NIAAA; Wood 2018

  • Laboratory range
  • Evidence supports
  • Guideline goal or unstudied
  • Higher risk in studies
Scale in standard drinks per week. A US standard drink is 14 grams (0.6 fluid ounces) of pure alcohol: 12 ounces of 5 percent beer, 5 ounces of 12 percent wine or 1.5 ounces of 80 proof spirits (US Dietary Guidelines 2020 to 2025). Other countries define it differently: Canada uses 13.45 grams, the Global Burden of Disease study 10 grams, and across 37 governments the definition ranges from 8 to 20 grams. Each source is quoted in its own unit. Bands show where studies found something, not a recommendation for any one person.

02 · What it is

What alcohol does.

Ethanol is a small, water-soluble molecule that the body cannot store and must oxidize. Alcohol dehydrogenase converts it first to acetaldehyde, a mutagen and animal carcinogen that damages DNA, and the mitochondrial enzyme ALDH2 then converts acetaldehyde to acetate; acetate, not fat, is the main fate of a drink. In eight healthy men given 24 grams of alcohol, 77 percent of the alcohol cleared from plasma was converted directly to acetate entering the blood, acetate turnover rose two and a half fold, the release of fatty acids from fat tissue fell by 53 percent and whole-body fat oxidation fell by 73 percent, while the share of newly made fat in the liver's triglyceride output rose from 2 to 30 percent. That is the metabolic signature of alcohol: the liver stops burning fat while it burns ethanol.1,2Grade C

The US Dietary Guidelines 2020 to 2025 define one alcoholic drink equivalent as 14 grams (0.6 fluid ounces) of pure alcohol: 12 fluid ounces of regular beer at 5 percent, 5 fluid ounces of wine at 12 percent, or 1.5 fluid ounces of 80 proof distilled spirits. Adults of legal drinking age can choose not to drink, or to drink in moderation by limiting intake to 2 drinks or less in a day for men and 1 drink or less in a day for women on days when alcohol is consumed; the guideline states that this is not an average over several days but the amount on any single day, that it does not recommend that people who do not drink start for any reason, and that drinking less is better for health than drinking more. The National Institute on Alcohol Abuse and Alcoholism defines binge drinking as a pattern that brings blood alcohol to 0.08 percent, about 5 drinks for a man or 4 for a woman in about two hours, and heavy drinking as 5 or more drinks on any day or 15 or more a week for men, and 4 or more on any day or 8 or more a week for women.3,4Grade A

What counts as a drink and as low risk differs by country. Among 37 governments with a definition, the modal standard drink was 10 grams of ethanol with a range of 8 to 20 grams, and low-risk guidelines ranged from 10 to 42 grams a day for women and 10 to 56 for men, or 98 to 140 grams a week for women and 150 to 280 for men. Canada's 2023 guidance uses a 13.45 gram drink and describes a continuum of risk: low at 2 standard drinks or less a week, moderate at 3 to 6, and increasingly high at 7 or more, with no more than 2 on any day. The 2016 Global Burden of Disease analysis counted in 10 gram drinks. A reader comparing guidelines is comparing different glasses.5,6,7Grade A

03 · Origins

Where it came from, and where it sits in the network.

Humans could metabolize ethanol long before anyone brewed it. Resurrecting the digestive alcohol dehydrogenase ADH4 of our primate ancestors in the laboratory showed that the common ancestor of humans and African apes gained an enzyme capable of oxidizing ethanol efficiently about 10 million years ago, near the time those apes began using the forest floor, where fallen fruit carries more fermenting yeast and more ethanol than fruit still on the tree; the ADH4 of earlier, tree-dwelling ancestors did not oxidize ethanol well. The frugivory hypothesis reads the attraction from the same history: ethanol in ripe and decaying fruit signals sugar, volatilized alcohols help locate ripe fruit by smell, and a taste for it would have been selected in any fruit-eating primate, so that modern patterns of drinking may be a maladaptive co-option of an ancestral nutritional strategy. Both are hypotheses with laboratory and comparative support, not settled fact.8,9Grade C

The second enzyme in the chain is where human populations differ most. A variant of the mitochondrial aldehyde dehydrogenase ALDH2 (ALDH2*504Lys, often written ALDH2*2) acts as a dominant negative, cutting activity more than a hundredfold in people with one copy and abolishing it in people with two; in data on 84,782 individuals from 452 population samples it is essentially absent everywhere except East Asia, is most frequent in southeast China and appears to have spread with the Han Chinese across China, Japan, Korea, Mongolia and Indochina. Carriers accumulate acetaldehyde when they drink and flush, feel nauseated and get a racing heart; an estimated 540 million people carry the variant. Those with two copies can barely drink and are protected from alcohol-related esophageal cancer, while those with one copy who drink through the flush carry odds of esophageal squamous cell carcinoma of 3.7 to 18.1 times in Japanese and Taiwanese case-control studies after adjustment for how much they drink. The flushing response is both a natural experiment that proved acetaldehyde carcinogenic and a reason the same number of drinks is not the same exposure in every body; it is also the genetic handle the China Kadoorie Biobank used to test the J-curve.10,1,11Grade C

On the shelf, ethanol sits upstream of a cluster of markers because the liver clears it ahead of everything else. While it does, fat oxidation nearly stops, fatty acids are redirected into triglyceride, the liver makes larger VLDL particles, and an acute block of lipoprotein lipase adds to the post-meal triglyceride peak when alcohol accompanies a fatty meal, which is why alcohol is a named cause of high triglycerides and, in people with diabetes or obesity, of the severe hypertriglyceridemia that can trigger pancreatitis. Ethanol perturbs almost every arm of hepatic lipid handling, and steatosis is a rapid response to it. GGT rises with regular intake and is read partly as a marker of the oxidative stress of ethanol metabolism; ALT rises with it, roughly linearly across 19,225 Finnish adults. Serum iron and ferritin rose progressively across classes of alcohol intake in 3,375 Australian twins, beer more than wine or spirits. Uric acid rises because ethanol speeds the turnover of adenine nucleotides: in six men with gout, eight days of oral ethanol raised serum urate from 8.4 to 10.1 mg/dL and whole-blood lactate from 1.3 to 3.1 mM. HDL rises, by a pooled 0.094 mmol/L in 44 intervention studies, which is the biomarker most often cited for a cardiac benefit. And blood pressure rises in step with intake, in self-reported drinkers and in genotype-predicted drinkers alike. One marker is a clue; this cluster moving together is a story.2,12,13,14,15,16,17,18,11Grade C

04 · The usual guidance

What the guidelines say.

No level the WHO calls safe; the US Dietary Guidelines 2020 to 2025 set a cap of 2 drinks or less in a day for men and 1 or less for women on days alcohol is consumed and do not recommend starting; Canada 2023 calls 2 standard drinks or less a week low risk, 3 to 6 moderate and 7 or more increasingly high.

The US Dietary Guidelines 2020 to 2025 state that among those who drink, higher average consumption is associated with higher death from all causes than lower average consumption; that emerging evidence suggests even drinking within the recommended limits may increase the overall risk of death from several cancers and some forms of cardiovascular disease; that alcohol raises cancer risk and for some cancers does so at less than 1 drink a day; and that caution is therefore recommended. The limits are 2 drinks or less in a day for men and 1 or less for women on days when alcohol is consumed, with binge drinking to be avoided. About 60 percent of US adults report drinking in the past month, and about 30 percent of those binge.3Grade A

In January 2023 the World Health Organization stated in The Lancet Public Health that alcohol is classified by the International Agency for Research on Cancer as a Group 1 carcinogen causally linked to seven cancers, including cancers of the esophagus, liver, colon and rectum, and breast; that in the European Union light to moderate drinking of less than 20 grams a day was associated with almost 23,000 new cancer cases in 2017, almost half of them breast cancers, and more than a third of them at less than 10 grams a day; that evidence does not indicate a threshold at which the carcinogenic effect begins; and that no safe amount of alcohol consumption for cancers and health can be established. The IARC classification itself dates from the 2007 and 2009 Monograph working groups.19,20Grade A

Canada's Guidance on Alcohol and Health, published in 2023 by the Canadian Centre on Substance Use and Addiction, replaced a daily limit with a weekly continuum: the risk is low for people who consume 2 standard drinks or less a week, moderate between 3 and 6, where the risk of several cancers including breast and colon cancer rises, and increasingly high at 7 or more, where the risk of heart disease and stroke rises significantly; whatever the weekly total, no more than 2 drinks on any day. The guidance makes the same recommendation for women and men, and states that not drinking has benefits such as better health and better sleep.6Grade A

05 · The evidence

What the studies support.

The largest pooled analysis of individual drinkers combined 83 prospective studies from 19 high-income countries: 599,912 current drinkers without prior cardiovascular disease, 40,310 deaths and 39,018 cardiovascular events over 5.4 million person-years, with hazard ratios corrected for the drift in people's drinking using 152,640 repeat assessments. All-cause mortality rose with intake in a positive, curvilinear way, with the minimum around or below 100 grams a week. Each 100 grams a week higher carried a hazard ratio of 1.14 for stroke, 1.06 for coronary disease other than heart attack, 1.09 for heart failure, 1.24 for fatal hypertensive disease and 1.15 for fatal aortic aneurysm, with no clear threshold below which less drinking stopped being associated with less disease; heart attack alone went the other way, 0.94 per 100 grams. Against drinkers at 100 grams a week or less, those at 100 to 200 grams had about six months less life expectancy at age 40, those at 200 to 350 grams one to two years less, and those above 350 grams four to five years less. The authors concluded that the data support limits lower than most current guidelines.21Grade B

The Global Burden of Disease 2016 analysis drew on 694 data sources of consumption and 592 studies of risk across 23 health outcomes. Alcohol was the seventh leading risk factor for death and disability in 2016, accounting for 2.2 percent of age-standardized female deaths and 6.8 percent of male deaths, and the leading risk factor among people aged 15 to 49, in whom it accounted for 3.8 percent of female and 12.2 percent of male deaths. Among people aged 50 and over, cancers made up 27.1 percent of alcohol-attributable female deaths and 18.9 percent of male deaths. Weighing the protective associations against the harms, the level of consumption that minimized overall health loss was zero standard drinks of 10 grams a week, with an uncertainty interval of 0.0 to 0.8. That is a modeled estimate across outcomes and populations, not a measurement in any one person.7Grade B

The J-curve is the central contested question, and both sides should be stated. The case for it: pooling 34 prospective studies with 1,015,835 people and 94,533 deaths, intake up to 4 drinks a day in men and 2 in women was inversely associated with total mortality, with a maximum protection of 17 percent in men and 18 percent in women; across 84 cohort studies, drinkers against non-drinkers had a relative risk of 0.75 for cardiovascular mortality, 0.71 for incident coronary disease and 0.87 for all-cause mortality, with the lowest coronary mortality at 1 to 2 drinks a day, and in 44 intervention studies alcohol raised HDL cholesterol, apolipoprotein A1 and adiponectin and lowered fibrinogen, which the authors read as pathophysiological support for a protective effect. Those are real, repeated observations.22,23,18Grade B

The case against it rests on who the abstainers were. In a meta-analysis of 87 studies with 3,998,626 people and 367,103 deaths, the unadjusted pooling reproduced the classic J, with low-volume drinkers at a relative risk of 0.86, but former drinkers carried a risk of 1.22, and after adjustment for misclassifying former and occasional drinkers as abstainers and for study quality, the low-volume benefit disappeared (0.97, 0.88 to 1.07); the higher-quality studies without those biases showed none either. The 2023 update of 107 cohort studies, 4,838,825 participants and 425,564 deaths found no significantly lower mortality in occasional drinkers (0.96) or low-volume drinkers of 1.3 to 24 grams a day (0.93, P = .07) than in lifetime non-drinkers, a non-significant rise at 25 to 44 grams a day (1.05), and significant rises at 45 to 64 and 65 or more grams a day (1.19 and 1.35), with risk beginning at lower intakes in women (1.22). In UK Biobank, light to moderate drinkers also had healthier habits, and adjusting for them attenuated the apparent protection before any genetics were applied.24,25,26Grade B

Mendelian randomization is the verdict, because a gene that changes how much a person drinks is not changed by their income, their diet or their early illness. The China Kadoorie Biobank enrolled 512,715 adults and genotyped 161,498 for two variants that alter alcohol metabolism, ALDH2-rs671 and ADH1B-rs1229984; 33 percent of men but only 2 percent of women drank in most weeks, so the women served as a control for the genes themselves. By self-report, men drinking about 100 grams a week had lower risks of ischemic stroke, brain hemorrhage and heart attack than non-drinkers or heavier drinkers, the familiar U. By genotype, with predicted intake ranging from 4 to 256 grams a week, there was no U at all: stroke risk rose continuously, with a relative risk per 280 grams a week of 1.58 for brain hemorrhage and 1.27 for ischemic stroke, heart attack risk did not fall (0.96, 0.78 to 1.18), and blood pressure rose with intake however it was measured. In women the genotypes did not predict drinking and were not associated with any of it. The authors' interpretation: the apparent protection of moderate drinking against stroke is largely non-causal, and alcohol uniformly increases blood pressure and stroke risk.11Grade B

European genetic studies point the same way. Across 56 studies and 261,991 people of European descent, carriers of the ADH1B variant that lowers drinking consumed 17.2 percent fewer units a week, binged less and abstained more, and had lower systolic pressure (0.88 mm Hg), interleukin-6, waist and body mass index, with lower odds of coronary heart disease (0.90) and ischemic stroke (0.83); the protection held across every category of consumption including the lightest drinkers, which is the opposite of what a true J-curve would predict. In UK Biobank, 371,463 people drinking a mean of 9.2 standard drinks a week, a one standard deviation increase in genetically predicted intake carried 1.3-fold higher odds of hypertension and 1.4-fold higher odds of coronary artery disease, and nonlinear analysis showed light intake adding little risk and heavier intake adding it exponentially; the authors read the genetics as alcohol at all amounts raising cardiovascular risk, with large differences across levels. Using up to 94 variants, genetically predicted drinking carried an odds ratio of 1.27 for stroke and 3.05 for peripheral artery disease per standard deviation, with borderline signals for coronary disease (1.16), atrial fibrillation (1.17) and abdominal aortic aneurysm (2.60), and nothing for heart failure, venous thrombosis or aortic stenosis.27,26,28Grade B

Cancer has the clearest dose-response. Across 572 studies and 486,538 cancer cases, heavy drinkers against non-drinkers and occasional drinkers had relative risks of 5.13 for oral and pharyngeal cancer, 4.95 for esophageal squamous cell carcinoma, 2.65 for laryngeal, 2.07 for liver, 1.61 for breast and 1.44 for colorectal cancer, each with a clear dose-risk relationship, and higher risks of stomach, gallbladder, pancreatic and lung cancer. In 2020 an estimated 741,300 new cancers worldwide, 4.1 percent of all cases, were attributable to alcohol; heavy drinking above 60 grams a day accounted for 46.7 percent of them, risky drinking of 20 to 60 grams for 39.4 percent, and moderate drinking below 20 grams a day for 103,100 cases, 13.9 percent, with 41,300 cases at up to 10 grams a day. For breast cancer the threshold is low: among 105,986 nurses followed from 1980 to 2008 with 7,690 invasive breast cancers, risk was significantly raised at 5.0 to 9.9 grams a day, the equivalent of 3 to 6 drinks a week (relative risk 1.15), binge drinking added risk independent of total intake, and intake earlier and later in adult life each counted.29,30,31Grade A

Atrial fibrillation rises from modest intakes. Among 79,019 Swedish adults with 7,245 new cases, relative to under 1 drink a week the risk was 1.14 at 15 to 21 drinks a week and 1.39 above 21 (12 gram drinks), and a meta-analysis of 7 prospective studies with 12,554 cases gave relative risks of 1.08 for 1 drink a day, 1.17 for 2, 1.26 for 3, 1.36 for 4 and 1.47 for 5 against non-drinkers. In 107,845 Europeans with a median intake of 3 grams a day followed a median of 13.9 years, with 5,854 new cases, the hazard rose non-linearly with intake and was 1.16 per 12 gram drink a day, similar across beverage types and not explained by cardiac biomarkers or by heart failure; the authors note that even 1.2 drinks a day carried higher risk. The one randomized trial of stopping enrolled 140 patients with paroxysmal or persistent atrial fibrillation who drank 10 or more standard drinks a week: the abstinence group cut intake from 16.8 to 2.1 drinks a week, the control group from 16.4 to 13.2, and over six months atrial fibrillation recurred in 53 percent of the abstainers against 73 percent of controls (hazard ratio 0.55), with less than half the burden of time in the arrhythmia (0.5 against 1.2 percent).32,33,34Grade B

Blood pressure is the marker with trial evidence. Across 36 trials and 2,865 participants, reducing alcohol did not significantly lower blood pressure in people who drank two or fewer drinks a day, but in people who drank more it did, in a dose-dependent way with an apparent threshold; those drinking six or more drinks a day who cut intake by about half lowered systolic pressure by 5.50 mm Hg and diastolic by 3.97. In 20 cohort studies with 361,254 participants and 90,160 new cases of hypertension, men drinking 1 to 2 drinks a day had a relative risk of 1.19 against abstainers, 1.51 at 3 to 4 and 1.74 at 5 or more, while women showed no increase at 1 to 2 drinks a day (0.94) and an increase beyond it (1.42); the authors found no evidence of a protective effect in women, contrary to earlier pooling. In the Chinese genetic study, blood pressure rose with predicted intake in men with no floor.35,36,11Grade A

Pattern matters independently of total. Across 14 studies with 4,718 ischemic heart disease events, people who averaged moderate intake but had irregular heavy occasions of more than 60 grams or 5 or more drinks at least monthly had a relative risk of 1.45 against regular moderate drinkers, and the authors concluded that the cardioprotective association of moderate drinking disappears when it is mixed with heavy occasions. For the liver, among 55,917 Danes aged 50 to 64 with 257 cirrhosis cases in men and 85 in women, and none among lifetime abstainers, daily drinking carried a hazard ratio of 3.65 against drinking on 2 to 4 days a week in men, recent intake predicted cirrhosis and intake in the twenties and thirties did not, and among men drinking 14 to 28 drinks a week the hazard was 7.47 for those drinking little wine and 1.69 for those drinking mostly wine against men drinking under 14. Among 19,225 Finnish adults, people whose overall intake was low-risk but who binged more than once a month had markedly higher GGT and ALT than those who never did.37,38,15Grade B

The brain and sleep carry smaller but consistent signals. In 550 Whitehall II civil servants with weekly intake and cognition measured over 30 years and brain MRI at the end, higher intake raised the odds of hippocampal atrophy in a dose-dependent way: 5.8 above 30 UK units a week against abstainers, and 3.4 for right-sided atrophy at 14 to 21 units a week, the range then within UK guidance, with no protective effect of light drinking of 1 to under 7 units a week over abstinence and a faster decline in lexical fluency. On sleep, a qualitative review of every published study in healthy volunteers found that at all doses alcohol shortens the time to fall asleep, consolidates the first half of the night and increases disruption in the second half, delays the first REM period at all doses and reduces total REM at moderate and high doses; the nightcap buys an easier start and a worse finish.39,40Grade C

06 · What moves it

The levers with evidence behind them.

What is listed here has been tested. The size of each effect, and the size of the study, are in the sentence, so a small effect reads as a small effect.

  • The month off: a markers test

    shows within four weeks what drinking is doing to a particular person's blood pressure, insulin resistance, weight and liver enzymes

    In a prospective study of 94 healthy adults drinking a mean of 258 grams a week who abstained for one month, against 47 who kept drinking a mean of 234 grams, the abstainers' insulin resistance score fell by 25.9 percent, systolic pressure by 6.6 percent, diastolic by 6.3 percent and weight by 1.5 percent, with falls of 41.8 and 73.9 percent in two cancer-related growth factors, none of it explained by changes in diet, exercise or smoking, and no change in the controls. GGT answers on its own clock: in 32 patients with alcohol-related liver disease who stayed abstinent for eight weeks, serum GGT fell with a half-life of 26 days, so a high GGT takes a month to halve and three to approach baseline. A month without alcohol, with the markers drawn before and after, is the cleanest test available of how much of a person's GGT, ALT, triglycerides, uric acid, ferritin and blood pressure belongs to the drinking.41,42Grade C

  • Pattern: drink-free days over the same total spread thin

    lowers cirrhosis and heart risk at a given weekly amount; heavy occasions undo the moderate average

    At the same weekly total, daily drinking carried a 3.65-fold hazard of cirrhosis against drinking on 2 to 4 days a week in Danish men, and irregular heavy occasions carried a relative risk of 1.45 for ischemic heart disease against regular moderate drinking across 14 studies. Binge drinking, not frequency, added breast cancer risk independent of total intake in 105,986 nurses, and binges more than once a month raised GGT and ALT even in people whose overall intake was low-risk. Canada's guidance makes the same point as a rule: whatever the weekly number, no more than 2 drinks on any day.38,37,31,15,6Grade B

  • Food with the drink

    lowers the peak blood alcohol and shortens the exposure; does not remove the dose

    In a crossover study, ten healthy men who drank 0.80 grams per kilogram of ethanol after an overnight fast reached a peak blood alcohol of 104 mg/dL; the same dose immediately after breakfast peaked at 67 mg/dL, with the area under the curve cut from 398 to 241 mg/dL times hours and the time to clear the dose about two hours shorter, and a meal eaten five hours after drinking still sped elimination by 36 to 50 percent. Food changes the shape of the exposure, not the grams of ethanol the liver must oxidize, and a fatty meal with alcohol adds to the post-meal triglyceride peak through an acute block of lipoprotein lipase.43,12Grade C

  • Medications that interact

    alcohol changes the drug and the drug changes the alcohol; some interactions occur at moderate intake

    Many medications interact with alcohol in two ways: pharmacokinetically, because alcohol and many drugs are metabolized in the liver by the same enzymes, and pharmacodynamically, because alcohol adds to a drug's effects, especially sedation. The classes named in the standard review include antibiotics, antidepressants, antihistamines, barbiturates, benzodiazepines, histamine H2 blockers, muscle relaxants, non-opioid pain relievers and anti-inflammatories, opioids and the blood thinner warfarin, along with many over-the-counter and herbal products, and some of these interactions occur at moderate drinking levels. GGT is also raised by enzyme-inducing drugs, which is one reason a high GGT is not a measure of drinking on its own. The US guideline lists people taking medications that can interact with alcohol among those who should not drink.44,45,3Grade C

  • The flushing response

    marks a person in whom the same drinks carry a much higher acetaldehyde load and cancer risk

    A person who flushes, feels sick or gets a racing heart after a drink is very likely carrying the low-activity ALDH2 variant, with more than a hundredfold less of the enzyme that clears acetaldehyde. In Japanese and Taiwanese studies, carriers of one copy who drank had odds of esophageal squamous cell carcinoma of 3.7 to 18.1 after adjustment for intake, most studies showing odds above 10 in heavy drinkers; the variant does not raise risk in non-drinkers. The authors' clinical advice is that physicians can identify these patients by asking about flushing and should counsel them to limit alcohol. The question costs nothing and changes the reading of every drink.1,10Grade C

  • What it does to GGT, ALT and the AST to ALT ratio

    raises GGT and ALT roughly in proportion to regular intake; a ratio of 2 or more points to established liver disease, not simply to heavy drinking

    GGT, AST, ALT and red-cell volume are the traditional markers of heavy drinking; they have imperfect sensitivity and specificity, are more likely to be raised in people over 30 and in regular drinkers with a long history, and must be read with obesity and hepatitis C in mind. Across 19,225 adults, GGT and ALT rose roughly linearly with the amount of regular drinking. In 313 patients admitted for alcohol withdrawal, the AST to ALT ratio was 1.0 or less in 64 percent and only exceptionally 2 or more, whereas in 48 patients admitted with alcohol-related cirrhosis 69 percent had a ratio of 2 or more; a high ratio suggests advanced liver disease rather than drinking as such. In the WHO/ISBRA study of 1,863 people in five countries, carbohydrate-deficient transferrin and GGT performed comparably as markers of high-risk drinking, both detected heavy drinking better than intermediate drinking, both were influenced by body mass index, sex, age and smoking, and because they are relatively independent their combination outperforms either alone.14,15,46,47Grade C

  • What it does to ferritin and iron

    raises ferritin and serum iron from low intakes; above 2 drinks a day, raises the prevalence of iron-overload markers

    Among 3,375 adult twins, serum iron and ferritin increased progressively across classes of alcohol intake, beer more than wine or spirits, ferritin was higher in people who had ever been alcohol dependent, and the effect did not depend on HFE genotype or body mass index; the authors concluded that even low-level intake raises ferritin and, by inference, body iron stores. In the US national survey, people drinking more than 2 drinks a day had a significantly higher prevalence of every marker of iron overload (transferrin saturation above 45 percent, ferritin above 300 ng/mL and their combinations) than non-drinkers after adjustment, while any alcohol was associated with a 40 percent lower risk of iron-deficiency anemia. A raised ferritin in someone who drinks is not inflammation or hemochromatosis until the drinking has been subtracted.16,48Grade C

  • What it does to triglycerides and HDL

    raises HDL in trials; raises triglycerides after fatty meals and with chronic excess, while fasting triglycerides in short trials did not move

    Across 44 intervention studies in adults without cardiovascular disease, alcohol raised HDL cholesterol by a pooled 0.094 mmol/L with a dose-response, raised apolipoprotein A1 and adiponectin, lowered fibrinogen, and did not change fasting triglycerides, with similar results for wine, beer and spirits. The triglyceride effect is post-prandial and chronic: alcohol with a meal containing fat adds to the triglyceride peak through acute inhibition of lipoprotein lipase, and with chronic excess the liver's output of large VLDL particles is the main source of the hypertriglyceridemia, which in people with type 2 diabetes or obesity can become severe enough to cause pancreatitis. The HDL rise is the biomarker most often cited for a cardiac benefit; the genetic studies above suggest it does not translate into fewer events.18,12,27Grade B

  • What it does to uric acid and gout

    raises urate production; doubles gout risk at 30 to 50 grams a day, beer most

    Among 47,150 men followed 12 years with 730 confirmed cases of gout, the relative risk against non-drinkers was 1.32 at 10 to 14.9 grams a day, 1.49 at 15 to 29.9, 1.96 at 30 to 49.9 and 2.53 at 50 or more; each daily 12 ounce beer carried 1.49, each shot of spirits 1.15, and a 4 ounce glass of wine 1.04, not significant. The mechanism is production, not just excretion: eight days of oral ethanol in six men with gout raised serum urate from 8.4 to 10.1 mg/dL, raised urinary oxypurines to 641 percent of baseline and daily uric acid turnover to 170 percent, through faster turnover of adenine nucleotides.49,17Grade B

  • Direct markers of recent drinking

    phosphatidylethanol reflects the past two weeks and carbohydrate-deficient transferrin the past weeks; both are approximate

    Whole-blood phosphatidylethanol (PEth) is formed only in the presence of ethanol. Among 40 outpatients in treatment it was the most sensitive marker of current regular drinking, detecting 43 percent of relapses on its own and a further 38 percent together with carbohydrate-deficient transferrin. In 36 outpatients giving drinking diaries and blood at three to four week intervals, PEth correlated best with the past two weeks' intake, and a change of about 20 grams a day moved it by about 0.10 micromol/L, with large scatter between people, so a single value supports only an approximate estimate and can misclassify moderate and heavy drinking. These are tools for when the account and the indirect markers disagree, not a routine panel.50,51,47Grade C

07 · Reading it well

Caveats, and what belongs with a physician.

Nearly all of the dose-response evidence is self-reported drinking at one point in time, and people under-report. The pooled analysis of 599,912 drinkers corrected for drift using repeat assessments, and the genetic studies sidestep the problem by instrumenting intake with ALDH2 and ADH1B variants, but the cancer and atrial fibrillation cohorts rest on questionnaires. The direction of the error is known: true intake is usually higher than reported, which places the risks at lower real doses than the categories suggest.21,11,32Grade B

The genetic verdict on the heart is not unanimous in detail. In the Chinese study, genotype-predicted intake had no effect on heart attack (0.96), in the European ADH1B pooling the low-drinking variant carried lower coronary disease (0.90), the 94-variant analysis gave a borderline 1.16, and the pooled drinkers showed a lower heart attack hazard with more drinking (0.94 per 100 grams) alongside higher stroke, heart failure and hypertensive death. The honest summary is that alcohol has little or no net benefit for the coronary arteries and clear harm for blood pressure, stroke and the atria; a null for heart attack is not a benefit.11,27,28,21Grade B

Population guidance is written in grams and weeks, and bodies differ. The China Kadoorie men drank mainly spirits and 98 percent of the women did not drink; the ALDH2 variant that made the natural experiment possible is essentially absent outside East Asia; breast cancer risk began at 3 to 6 drinks a week in women, and the hypertension pooling found risk from 1 to 2 drinks a day in men but not women. The bands on this page quote each source in its own unit, converted to US drinks only where the label says so, and no single number on it is a personal threshold.11,10,31,36Grade B

The markers alcohol raises are also raised by other things. GGT rises with obstructive liver disease, obesity and enzyme-inducing drugs; ALT with fatty liver, viral hepatitis and a week of hard lifting; ferritin with inflammation and iron overload; triglycerides with sugar and insulin resistance. In the WHO/ISBRA sample, body mass index, sex, age and smoking all shifted GGT and CDT, and the traditional markers have imperfect sensitivity and specificity for drinking. A pattern across several markers, with the drinking history and a month-off test, carries more weight than any one value.45,14,47Grade C

The one-month abstinence study was observational and self-selected: the 94 abstainers chose to stop and the 47 controls chose not to, and the exclusion of anyone with alcohol dependence or known liver disease means its figures describe moderate-to-heavy social drinkers, not people who cannot stop. The atrial fibrillation trial was open-label, 85 percent men, and six months long.41,34Grade C

See a physician

  • Anyone who drinks heavily every day should not stop abruptly without medical advice. Withdrawal can begin within hours of the last drink with tremor, sweating, a racing heart, anxiety and insomnia, and in a minority progresses to seizures or delirium, which are medical emergencies; a physician can judge the risk and supervise the stop.
  • Jaundice, a swollen abdomen, vomiting blood or black stools, easy bruising or confusion in someone who drinks, which can mean advanced liver disease.
  • An AST to ALT ratio of 2 or more, a GGT that stays high after a month off, a falling platelet count or a rising bilirubin, which call for a liver assessment rather than another month of watching.
  • Palpitations, an irregular pulse or a diagnosis of atrial fibrillation in someone who drinks regularly; stopping halved recurrence in the one randomized trial and belongs in the treatment plan.
  • Drinking during pregnancy or while trying to conceive, drinking with medications that interact, or drinking with a history of pancreatitis or very high triglycerides.
  • Wanting to cut down and finding you cannot, drinking in the morning, blackouts, or others expressing concern; alcohol use disorder is a medical condition with effective treatment, and the conversation is confidential.
  • Flushing, nausea or a racing heart after small amounts of alcohol, especially with East Asian ancestry, together with any difficulty swallowing, which should prompt a lower limit and attention to the esophagus.

This page is education, not individual medical advice, and reading it creates no physician-patient relationship.

08 · Open questions

What the literature does not settle.

  • No randomized trial has tested moderate drinking against abstinence for cardiovascular events or mortality; a large US trial was planned and cancelled, and the causal evidence therefore rests on Mendelian randomization, which tests lifelong genetic differences in intake rather than a decision made at 50.
  • Whether the small inverse association with heart attack that survives in some analyses is real, and whether it is outweighed by stroke, atrial fibrillation and cancer in every population, is still argued; the net estimate depends on which outcomes are counted and how former drinkers are classified.
  • The dose-response below about one drink a day is imprecise in every design: the cohorts have few lifetime abstainers and the genetic instruments predict small differences in intake, so the shape of the curve between zero and seven drinks a week is estimated, not measured.
  • Beverage type is unsettled. Wine carried less cirrhosis risk than beer and spirits at the same intake in the Danish cohort and no gout risk in the US men, while atrial fibrillation risk was similar across types; whether this is the drink, the drinker or the pattern is not resolved.
  • How fast each marker recovers in a light or moderate drinker who stops is poorly documented. The GGT half-life of 26 days comes from patients with liver disease, and the one-month abstinence data come from moderate-to-heavy drinkers; the recovery curves of ferritin, uric acid and triglycerides after a month off in an ordinary drinker have not been mapped.
  • Whether the brain findings in Whitehall II, hippocampal atrophy at 14 to 21 UK units a week, translate into dementia at those intakes is not established; the study found no association with cross-sectional cognitive performance.

09 · In practice

How Dr. Tagge reads it in his own practice.

I do not moralize about alcohol and I do not hand out a number of my own; the figures on this page are the guidelines' and the cohorts', and I quote them as theirs. What I do is read drinking through the markers, against each other, because one marker is a clue and three are a story. A GGT that is up, an ALT that is up, a ferritin that is up and a triglyceride that is up, in a person whose uric acid is edging higher and whose blood pressure has crept, is a pattern I have learned to ask about before I ask about anything else, and I ask about the pattern, not just the total: how many days a week, how many on the biggest day, whether it comes with dinner or after it, whether it has changed, and whether you flush. Then I look at sleep, because the second half of the night is where a drink shows first, and at the morning resting heart rate on a wearable if there is one. The markers answer fast. GGT halves in about a month, blood pressure and insulin resistance move measurably within four weeks of stopping in people who drink in the moderate-to-heavy range, and ferritin and triglycerides follow. So when the story points at alcohol, I do not argue about it; I ask for a month off with the same panel drawn before and after, and we let the numbers settle the question. Sometimes they clear the drinking and we look elsewhere. I will say plainly where I am ahead of the evidence. The genetic studies persuade me that the J-curve was mostly an artifact of who the abstainers were, and I act as though a daily drink buys nothing for the heart; no trial has randomized moderate drinkers to stop and counted heart attacks, so that is a reading of the evidence, not a result. I never tell anyone to start drinking, and I never tell anyone who drinks heavily every day to stop suddenly on their own, because withdrawal is a medical event and we plan it together.

A practice judgment, labeled as one (Grade D): it is how one physician reads the number for the people he cares for, not a recommendation for you.

10

Questions

11 · Sources

51 sources, read in full.

Numbered in the order they appear. Each line says what this page relies on from the paper; the link opens the record at PubMed or the publisher.

  1. 1.

    Brooks PJ, Enoch MA, Goldman D, Li TK, Yokoyama A. The alcohol flushing response: an unrecognized risk factor for esophageal cancer from alcohol consumption. PLoS Medicine. 2009. PMID 19320537. DOI 10.1371/journal.pmed.1000050.

    Review · ALDH2 deficiency (about 540 million people) causes flushing, nausea and tachycardia from acetaldehyde, a mutagen and animal carcinogen; heterozygotes have more than a hundredfold less enzyme activity and, when they drink, odds ratios of 3.7 to 18.1 for esophageal squamous cell carcinoma after adjustment for intake; the variant does not raise risk in non-drinkers.

  2. 2.

    Siler SQ, Neese RA, Hellerstein MK. De novo lipogenesis, lipid kinetics, and whole-body lipid balances in humans after acute alcohol consumption. American Journal of Clinical Nutrition. 1999. PMID 10539756. DOI 10.1093/ajcn/70.5.928.

    Laboratory method study · In eight healthy men given 24 g of alcohol, the fractional contribution of de novo lipogenesis to VLDL triglyceride rose from 2 to 30 percent but the absolute amount was under 5 percent of the dose; 77 percent of cleared alcohol became plasma acetate, acetate flux rose 2.5-fold, adipose fatty acid release fell 53 percent and whole-body lipid oxidation fell 73 percent.

  3. 3.

    US Department of Agriculture and US Department of Health and Human Services. Dietary Guidelines for Americans, 2020-2025, 9th edition: Alcoholic Beverages. US Department of Agriculture. 2020.

    Guideline · One drink equivalent is 14 g (0.6 fl oz) of pure alcohol; adults who drink should limit intake to 2 drinks or less in a day for men and 1 or less for women on days alcohol is consumed, not as an average; the guideline does not recommend that non-drinkers start, states that drinking less is better than drinking more, and notes emerging evidence of risk even within the limits and cancer risk at under 1 drink a day.

  4. 4.

    National Institute on Alcohol Abuse and Alcoholism. Drinking Levels and Patterns Defined. National Institutes of Health. 2024.

    Guideline · A US standard drink contains 0.6 fl oz or 14 g of pure alcohol; binge drinking brings blood alcohol to 0.08 percent, typically 5 or more drinks for a man or 4 or more for a woman in about two hours; heavy drinking is 5 or more on any day or 15 or more a week for men, and 4 or more on any day or 8 or more a week for women.

  5. 5.

    Kalinowski A, Humphreys K. Governmental standard drink definitions and low-risk alcohol consumption guidelines in 37 countries. Addiction. 2016. PMID 27073140. DOI 10.1111/add.13341.

    Review · Among 37 governments with a definition, the modal standard drink was 10 g of ethanol with a range of 8 to 20 g; low-risk guidelines ranged from 10 to 42 g a day for women and 10 to 56 g for men, or 98 to 140 g a week for women and 150 to 280 g for men.

  6. 6.

    Paradis C, Butt P, Shield K, et al.; Low-Risk Alcohol Drinking Guidelines Scientific Expert Panels. Canada's Guidance on Alcohol and Health: Final Report. Canadian Centre on Substance Use and Addiction. 2023.

    Guideline · A Canadian standard drink is 17.05 mL or 13.45 g of pure alcohol; the risk of harm is low at 2 standard drinks or less a week, moderate at 3 to 6 (several cancers including breast and colon rise) and increasingly high at 7 or more (heart disease and stroke rise significantly), with no more than 2 drinks on any day and the same guidance for women and men.

  7. 7.

    GBD 2016 Alcohol Collaborators. Alcohol use and burden for 195 countries and territories, 1990-2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet. 2018. PMID 30146330. DOI 10.1016/S0140-6736(18)31310-2.

    Systematic review and meta-analysis · From 694 consumption sources and 592 risk studies across 23 outcomes, alcohol was the seventh leading risk factor for death and disability in 2016 (2.2 percent of female and 6.8 percent of male deaths) and the leading one at ages 15 to 49; the level of consumption that minimized overall health loss was zero standard drinks (10 g) a week (95 percent UI 0.0 to 0.8).

  8. 8.

    Carrigan MA, Uryasev O, Frye CB, et al.. Hominids adapted to metabolize ethanol long before human-directed fermentation. Proceedings of the National Academy of Sciences of the United States of America. 2015. PMID 25453080. DOI 10.1073/pnas.1404167111.

    Laboratory method study · Resurrected ancestral digestive alcohol dehydrogenases (ADH4) show that our ape ancestors gained an enzyme capable of metabolizing ethanol about 10 million years ago, near the time they began using the forest floor, where fallen fruit carries more fermenting yeast and ethanol; earlier arboreal ancestors' ADH4 did not oxidize ethanol efficiently.

  9. 9.

    Dudley R. Evolutionary origins of human alcoholism in primate frugivory. Quarterly Review of Biology. 2000. PMID 10721531. DOI 10.1086/393255.

    Review · Ethanol in ripe and decaying fruit, olfactory localization of ripe fruit by volatilized alcohols and ethanol as an appetitive stimulant likely characterized all frugivorous primates including the human lineage; modern drinking may be a maladaptive co-option of an ancestral nutritional strategy.

  10. 10.

    Li H, Borinskaya S, Yoshimura K, et al.. Refined geographic distribution of the oriental ALDH2*504Lys (nee 487Lys) variant. Annals of Human Genetics. 2009. PMID 19456322. DOI 10.1111/j.1469-1809.2009.00517.x.

    Cross-sectional study · The ALDH2*504Lys variant acts as a dominant negative, greatly reducing activity in heterozygotes and abolishing it in homozygotes; in 4,091 newly typed and 80,691 published individuals from 452 population samples it is essentially absent outside East Asia, most frequent in southeast China, and appears to have spread with the Han Chinese.

  11. 11.

    Millwood IY, Walters RG, Mei XW, et al.; China Kadoorie Biobank Collaborative Group. Conventional and genetic evidence on alcohol and vascular disease aetiology: a prospective study of 500 000 men and women in China. Lancet. 2019. PMID 30955975. DOI 10.1016/S0140-6736(18)31772-0.

    Mendelian randomization · In 512,715 adults with 161,498 genotyped for ALDH2-rs671 and ADH1B-rs1229984, self-reported intake showed U-shaped associations with ischemic stroke, brain hemorrhage and MI, but genotype-predicted intake (4 to 256 g a week) showed none: stroke risk rose log-linearly (RR per 280 g a week 1.58 for hemorrhage, 1.27 for ischemic stroke), MI did not change (0.96), and blood pressure rose with intake; the apparent protection of moderate drinking is largely non-causal.

  12. 12.

    Van de Wiel A. The effect of alcohol on postprandial and fasting triglycerides. International Journal of Vascular Medicine. 2012. PMID 21961068. DOI 10.1155/2012/862504.

    Review · Alcohol with a fat-containing meal adds to the postprandial triglyceride peak through acute inhibition of lipoprotein lipase; with chronic excess, increased hepatic synthesis of large VLDL is the main source of hypertriglyceridemia, and in type 2 diabetes or obesity it can become severe enough to cause pancreatitis.

  13. 13.

    You M, Arteel GE. Effect of ethanol on lipid metabolism. Journal of Hepatology. 2019. PMID 30658725. DOI 10.1016/j.jhep.2018.10.037.

    Review · Ethanol perturbs almost all aspects of hepatic lipid metabolism, partly through the large metabolic demand alcohol oxidation places on the liver, so that steatosis is a rapid response to alcohol; transient steatosis is largely inert but chronic alcohol-related liver disease seems to require it.

  14. 14.

    Conigrave KM, Davies P, Haber P, Whitfield JB. Traditional markers of excessive alcohol use. Addiction. 2003. PMID 14984240. DOI 10.1046/j.1359-6357.2003.00581.x.

    Review · GGT, AST, ALT and mean corpuscular volume have imperfect sensitivity and specificity for excessive drinking, are more often raised in people over 30 and in regular long-term drinkers, GGT may partly mark the oxidative stress of ethanol metabolism, and obesity and hepatitis C must be considered when reading them.

  15. 15.

    Nivukoski U, Bloigu A, Bloigu R, Aalto M, Laatikainen T, Niemelä O. Liver enzymes in alcohol consumers with or without binge drinking. Alcohol. 2019. PMID 30890357. DOI 10.1016/j.alcohol.2019.03.001.

    Cross-sectional study · In 19,225 Finnish adults aged 25 to 74, GGT and ALT rose roughly linearly with regular intake; among people with low-risk overall consumption, those who binged more than once a month had markedly higher GGT and ALT than those who never did.

  16. 16.

    Whitfield JB, Zhu G, Heath AC, Powell LW, Martin NG. Effects of alcohol consumption on indices of iron stores and of iron stores on alcohol intake markers. Alcoholism: Clinical and Experimental Research. 2001. PMID 11505030.

    Cross-sectional study · In 3,375 adult twins, serum iron and ferritin rose progressively across classes of alcohol intake, beer more than wine or spirits, ferritin was higher in those ever alcohol dependent, there was no interaction with HFE genotype or BMI, and GGT, AST and ALT were higher in people with high ferritin.

  17. 17.

    Faller J, Fox IH. Ethanol-induced hyperuricemia: evidence for increased urate production by activation of adenine nucleotide turnover. New England Journal of Medicine. 1982. PMID 7144847. DOI 10.1056/NEJM198212233072602.

    Laboratory method study · In six patients with gout, oral ethanol (1.8 g/kg a day) for eight days raised serum urate from 8.4 to 10.1 mg/dL, whole-blood lactate from 1.3 to 3.1 mM, urinary oxypurines to 641 percent and daily uric acid turnover to 170 percent of baseline; ethanol increases urate synthesis by speeding adenine nucleotide turnover.

  18. 18.

    Brien SE, Ronksley PE, Turner BJ, Mukamal KJ, Ghali WA. Effect of alcohol consumption on biological markers associated with risk of coronary heart disease: systematic review and meta-analysis of interventional studies. BMJ. 2011. PMID 21343206. DOI 10.1136/bmj.d636.

    Systematic review and meta-analysis · Across 44 meta-analyzed intervention studies on 13 biomarkers in adults without cardiovascular disease, alcohol raised HDL cholesterol (0.094 mmol/L, dose-response), apolipoprotein A1 (0.101 g/L) and adiponectin (0.56 mg/L), lowered fibrinogen (0.20 g/L) and did not change triglycerides, similarly across wine, beer and spirits.

  19. 19.

    Anderson BO, Berdzuli N, Ilbawi A, et al.. Health and cancer risks associated with low levels of alcohol consumption. Lancet Public Health. 2023. PMID 36603913. DOI 10.1016/S2468-2667(22)00317-6.

    Consensus statement · WHO staff statement: alcohol is an IARC Group 1 carcinogen causally linked to seven cancers; in the EU, drinking under 20 g a day was associated with almost 23,000 new cancers in 2017, almost half breast cancers and more than a third at under 10 g a day; no threshold for the carcinogenic effect is evident, so no safe amount can be established.

  20. 20.

    Secretan B, Straif K, Baan R, et al.; WHO International Agency for Research on Cancer Monograph Working Group. A review of human carcinogens--Part E: tobacco, areca nut, alcohol, coal smoke, and salted fish. Lancet Oncology. 2009. PMID 19891056. DOI 10.1016/s1470-2045(09)70326-2.

    Consensus statement · The IARC Monograph Volume 100E working group's summary of its evaluation of alcoholic beverages as carcinogenic to humans (Group 1); the summary has no abstract and its classification wording is quoted here through the WHO 2023 statement, which cites it.

  21. 21.

    Wood AM, Kaptoge S, Butterworth AS, et al.; Emerging Risk Factors Collaboration/EPIC-CVD/UK Biobank Alcohol Study Group. Risk thresholds for alcohol consumption: combined analysis of individual-participant data for 599 912 current drinkers in 83 prospective studies. Lancet. 2018. PMID 29676281. DOI 10.1016/S0140-6736(18)30134-X.

    Systematic review and meta-analysis · In 599,912 current drinkers (40,310 deaths, 39,018 cardiovascular events), all-cause mortality was lowest around or below 100 g a week; per 100 g a week, HR 1.14 for stroke, 1.06 coronary disease excluding MI, 1.09 heart failure, 1.24 fatal hypertensive disease, 1.15 fatal aortic aneurysm, 0.94 for MI; life expectancy at 40 about 6 months, 1 to 2 years and 4 to 5 years shorter at 100 to 200, 200 to 350 and over 350 g a week.

  22. 22.

    Di Castelnuovo A, Costanzo S, Bagnardi V, Donati MB, Iacoviello L, de Gaetano G. Alcohol dosing and total mortality in men and women: an updated meta-analysis of 34 prospective studies. Archives of Internal Medicine. 2006. PMID 17159008. DOI 10.1001/archinte.166.22.2437.

    Systematic review and meta-analysis · Across 34 prospective studies (1,015,835 subjects, 94,533 deaths), a J-shaped association with total mortality: intake up to 4 drinks a day in men and 2 in women was inversely associated, with maximum protection of 17 percent in men and 18 percent in women; higher doses raised mortality.

  23. 23.

    Ronksley PE, Brien SE, Turner BJ, Mukamal KJ, Ghali WA. Association of alcohol consumption with selected cardiovascular disease outcomes: a systematic review and meta-analysis. BMJ. 2011. PMID 21343207. DOI 10.1136/bmj.d671.

    Systematic review and meta-analysis · Across 84 prospective studies, drinkers against non-drinkers had pooled RR 0.75 for cardiovascular mortality, 0.71 for incident coronary disease, 0.75 for coronary mortality, 0.98 for incident stroke, 1.06 for stroke mortality and 0.87 for all-cause mortality; lowest coronary mortality at 1 to 2 drinks a day.

  24. 24.

    Stockwell T, Zhao J, Panwar S, Roemer A, Naimi T, Chikritzhs T. Do “Moderate” Drinkers Have Reduced Mortality Risk? A Systematic Review and Meta-Analysis of Alcohol Consumption and All-Cause Mortality. Journal of Studies on Alcohol and Drugs. 2016. PMID 26997174. DOI 10.15288/jsad.2016.77.185.

    Systematic review and meta-analysis · Across 87 studies (3,998,626 people, 367,103 deaths), the unadjusted pooling reproduced the J-curve (low-volume RR 0.86; former drinkers 1.22), but after adjustment for abstainer bias and study quality low-volume drinkers had no reduced mortality (RR 0.97, 0.88 to 1.07), nor did higher-quality bias-free studies show one.

  25. 25.

    Zhao J, Stockwell T, Naimi T, Churchill S, Clay J, Sherk A. Association Between Daily Alcohol Intake and Risk of All-Cause Mortality: A Systematic Review and Meta-analyses. JAMA Network Open. 2023. PMID 37000449. DOI 10.1001/jamanetworkopen.2023.6185.

    Systematic review and meta-analysis · Across 107 cohort studies (4,838,825 participants, 425,564 deaths), after adjustment for former-drinker bias and study quality, occasional (RR 0.96) and low-volume drinkers of 1.3 to 24 g a day (0.93, P = .07) had no significantly lower mortality than lifetime non-drinkers; risk rose at 45 to 64 g a day (1.19) and 65 or more (1.35), and earlier in women (1.22).

  26. 26.

    Biddinger KJ, Emdin CA, Haas ME, et al.. Association of Habitual Alcohol Intake With Risk of Cardiovascular Disease. JAMA Network Open. 2022. PMID 35333364. DOI 10.1001/jamanetworkopen.2022.3849.

    Mendelian randomization · In 371,463 UK Biobank participants drinking a mean 9.2 standard drinks a week, healthier habits among light drinkers attenuated the observational protection; a 1 SD increase in genetically predicted intake carried 1.3-fold higher hypertension and 1.4-fold higher coronary artery disease, and nonlinear MR showed minimal added risk at light intake and exponential increases at heavier intake.

  27. 27.

    Holmes MV, Dale CE, Zuccolo L, et al.; InterAct Consortium. Association between alcohol and cardiovascular disease: Mendelian randomisation analysis based on individual participant data. BMJ. 2014. PMID 25011450. DOI 10.1136/bmj.g4164.

    Mendelian randomization · In 56 studies of 261,991 people of European descent, carriers of the ADH1B rs1229984 A-allele drank 17.2 percent fewer units a week, binged less and abstained more, had lower systolic pressure (0.88 mm Hg), interleukin-6, waist and BMI, and lower odds of coronary heart disease (0.90) and ischemic stroke (0.83), consistent across all categories of consumption.

  28. 28.

    Larsson SC, Burgess S, Mason AM, Michaëlsson K. Alcohol Consumption and Cardiovascular Disease: A Mendelian Randomization Study. Circulation: Genomic and Precision Medicine. 2020. PMID 32367730. DOI 10.1161/CIRCGEN.119.002814.

    Mendelian randomization · Using up to 94 variants, genetically predicted alcohol consumption carried OR 1.27 for stroke and 3.05 for peripheral artery disease per SD of log drinks a week, borderline associations with coronary artery disease (1.16), atrial fibrillation (1.17) and abdominal aortic aneurysm (2.60), and none with heart failure, venous thromboembolism or aortic stenosis.

  29. 29.

    Bagnardi V, Rota M, Botteri E, et al.. Alcohol consumption and site-specific cancer risk: a comprehensive dose-response meta-analysis. British Journal of Cancer. 2015. PMID 25422909. DOI 10.1038/bjc.2014.579.

    Systematic review and meta-analysis · Across 572 studies (486,538 cancer cases), heavy drinkers against non- and occasional drinkers had RR 5.13 for oral and pharyngeal cancer, 4.95 esophageal squamous cell, 2.65 laryngeal, 2.07 liver, 1.61 breast and 1.44 colorectal, each with a clear dose-risk relationship, plus higher stomach, gallbladder, pancreas and lung cancer.

  30. 30.

    Rumgay H, Shield K, Charvat H, et al.. Global burden of cancer in 2020 attributable to alcohol consumption: a population-based study. Lancet Oncology. 2021. PMID 34270924. DOI 10.1016/S1470-2045(21)00279-5.

    Cross-sectional study · An estimated 741,300 new cancers in 2020 (4.1 percent of all) were attributable to alcohol, led by esophagus, liver and breast; heavy drinking (over 60 g a day) accounted for 46.7 percent, risky drinking (20 to 60 g) 39.4 percent, moderate drinking (under 20 g) 103,100 cases (13.9 percent) and up to 10 g a day 41,300 cases.

  31. 31.

    Chen WY, Rosner B, Hankinson SE, Colditz GA, Willett WC. Moderate alcohol consumption during adult life, drinking patterns, and breast cancer risk. JAMA. 2011. PMID 22045766. DOI 10.1001/jama.2011.1590.

    Prospective cohort · Among 105,986 Nurses' Health Study women followed 1980 to 2008 (7,690 invasive breast cancers), risk was significantly raised at 5.0 to 9.9 g a day, equivalent to 3 to 6 drinks a week (RR 1.15); binge drinking but not frequency added risk after adjustment for cumulative intake, and intake earlier and later in adult life each counted.

  32. 32.

    Larsson SC, Drca N, Wolk A. Alcohol consumption and risk of atrial fibrillation: a prospective study and dose-response meta-analysis. Journal of the American College of Cardiology. 2014. PMID 25034065. DOI 10.1016/j.jacc.2014.03.048.

    Systematic review and meta-analysis · In 79,019 Swedish adults (7,245 incident AF), RR 1.14 at 15 to 21 drinks a week and 1.39 above 21 (12 g drinks) against under 1 a week; in a meta-analysis of 7 prospective studies (12,554 cases), RR 1.08, 1.17, 1.26, 1.36 and 1.47 for 1 to 5 drinks a day against non-drinkers.

  33. 33.

    Csengeri D, Sprünker NA, Di Castelnuovo A, et al.. Alcohol consumption, cardiac biomarkers, and risk of atrial fibrillation and adverse outcomes. European Heart Journal. 2021. PMID 33438022. DOI 10.1093/eurheartj/ehaa953.

    Prospective cohort · In 107,845 Europeans (median intake 3 g a day, median follow-up 13.9 years, 5,854 incident AF), alcohol was non-linearly and positively associated with AF, HR 1.16 per 12 g drink a day, similar across beverage types and not explained by cardiac biomarkers or heart failure; even 1.2 drinks a day carried higher risk.

  34. 34.

    Voskoboinik A, Kalman JM, De Silva A, et al.. Alcohol Abstinence in Drinkers with Atrial Fibrillation. New England Journal of Medicine. 2020. PMID 31893513. DOI 10.1056/NEJMoa1817591.

    Randomized trial · In 140 patients with paroxysmal or persistent AF drinking 10 or more standard drinks (12 g) a week, randomized abstinence cut intake from 16.8 to 2.1 drinks a week (controls 16.4 to 13.2); AF recurred in 53 percent against 73 percent over 6 months (HR 0.55) with lower AF burden (0.5 against 1.2 percent of time).

  35. 35.

    Roerecke M, Kaczorowski J, Tobe SW, Gmel G, Hasan OSM, Rehm J. The effect of a reduction in alcohol consumption on blood pressure: a systematic review and meta-analysis. Lancet Public Health. 2017. PMID 29253389. DOI 10.1016/S2468-2667(17)30003-8.

    Systematic review and meta-analysis · Across 36 trials (2,865 participants), reducing alcohol did not significantly lower blood pressure in people drinking 2 or fewer drinks a day but did in those drinking more, dose-dependently; those drinking 6 or more a day who halved intake lowered systolic pressure 5.50 mm Hg and diastolic 3.97.

  36. 36.

    Roerecke M, Tobe SW, Kaczorowski J, et al.. Sex-Specific Associations Between Alcohol Consumption and Incidence of Hypertension: A Systematic Review and Meta-Analysis of Cohort Studies. Journal of the American Heart Association. 2018. PMID 29950485. DOI 10.1161/JAHA.117.008202.

    Systematic review and meta-analysis · Across 20 cohort studies (361,254 participants, 90,160 incident hypertension), men drinking 1 to 2, 3 to 4 and 5 or more drinks a day (12 g) had RR 1.19, 1.51 and 1.74 against abstainers; women had no increase at 1 to 2 (0.94) and RR 1.42 above it; no protective effect in women.

  37. 37.

    Roerecke M, Rehm J. Irregular heavy drinking occasions and risk of ischemic heart disease: a systematic review and meta-analysis. American Journal of Epidemiology. 2010. PMID 20142394. DOI 10.1093/aje/kwp451.

    Systematic review and meta-analysis · Across 14 studies (4,718 ischemic heart disease events), irregular heavy drinking occasions (over 60 g or 5 or more drinks at least monthly) against regular moderate drinking carried pooled RR 1.45; the cardioprotective association of moderate drinking disappears when mixed with heavy occasions.

  38. 38.

    Askgaard G, Grønbæk M, Kjær MS, Tjønneland A, Tolstrup JS. Alcohol drinking pattern and risk of alcoholic liver cirrhosis: a prospective cohort study. Journal of Hepatology. 2015. PMID 25634330. DOI 10.1016/j.jhep.2014.12.005.

    Prospective cohort · Among 55,917 Danes aged 50 to 64 (257 cirrhosis cases in men, 85 in women, none in lifetime abstainers), daily drinking carried HR 3.65 against 2 to 4 days a week in men, recent rather than early-life intake predicted cirrhosis, and at 14 to 28 drinks a week the hazard was 7.47 for little wine and 1.69 for mostly wine against under 14.

  39. 39.

    Topiwala A, Allan CL, Valkanova V, et al.. Moderate alcohol consumption as risk factor for adverse brain outcomes and cognitive decline: longitudinal cohort study. BMJ. 2017. PMID 28588063. DOI 10.1136/bmj.j2353.

    Prospective cohort · In 550 Whitehall II adults (mean age 43 at baseline) followed 30 years with MRI at the end, higher intake raised the odds of hippocampal atrophy dose-dependently (5.8 above 30 UK units a week; 3.4 for right-sided atrophy at 14 to 21 units), with no protective effect of light drinking (1 to under 7 units) over abstinence and faster decline in lexical fluency.

  40. 40.

    Ebrahim IO, Shapiro CM, Williams AJ, Fenwick PB. Alcohol and sleep I: effects on normal sleep. Alcoholism: Clinical and Experimental Research. 2013. PMID 23347102. DOI 10.1111/acer.12006.

    Review · At all doses alcohol shortens sleep onset, consolidates the first half of the night and disrupts the second; it delays the first REM period at all doses, reduces total REM at moderate and high doses, and increases slow-wave sleep in the first half.

  41. 41.

    Mehta G, Macdonald S, Cronberg A, et al.. Short-term abstinence from alcohol and changes in cardiovascular risk factors, liver function tests and cancer-related growth factors: a prospective observational study. BMJ Open. 2018. PMID 29730627. DOI 10.1136/bmjopen-2017-020673.

    Prospective cohort · In 94 healthy moderate-to-heavy drinkers (mean 258 g a week) who abstained for one month against 47 who continued, HOMA insulin resistance fell 25.9 percent, systolic pressure 6.6 percent, diastolic 6.3 percent, weight 1.5 percent, VEGF 41.8 percent and EGF 73.9 percent, unrelated to changes in diet, exercise or smoking; controls did not change.

  42. 42.

    Orrego H, Blake JE, Israel Y. Relationship between gamma-glutamyl transpeptidase and mean urinary alcohol levels in alcoholics while drinking and after alcohol withdrawal. Alcoholism: Clinical and Experimental Research. 1985. PMID 2859812. DOI 10.1111/j.1530-0277.1985.tb05038.x.

    Prospective cohort · In 42 patients with alcohol-related liver disease, serum GGT correlated with mean urinary alcohol (r = 0.69); in 32 who stayed abstinent for 8 weeks, GGT fell with a half-life of 26 days.

  43. 43.

    Jones AW, Jönsson KA. Food-induced lowering of blood-ethanol profiles and increased rate of elimination immediately after a meal. Journal of Forensic Sciences. 1994. PMID 8064267.

    Randomized trial · In a two-part crossover study, ten healthy men drinking 0.80 g/kg ethanol fasting reached a peak blood alcohol of 104 mg/dL against 67 mg/dL after breakfast, with the area under the curve 398 against 241 mg/dL times hours and clearance about two hours faster; a meal five hours after drinking raised the elimination rate by 36 to 50 percent.

  44. 44.

    Weathermon R, Crabb DW. Alcohol and medication interactions. Alcohol Research and Health. 1999. PMID 10890797.

    Review · Alcohol interacts with medications pharmacokinetically (shared hepatic enzymes) and pharmacodynamically (added sedation), some interactions occurring at moderate intake; classes include antibiotics, antidepressants, antihistamines, barbiturates, benzodiazepines, H2 blockers, muscle relaxants, non-opioid pain and anti-inflammatory drugs, opioids and warfarin, plus many over-the-counter and herbal products.

  45. 45.

    Whitfield JB. Gamma glutamyl transferase. Critical Reviews in Clinical Laboratory Sciences. 2001. PMID 11563810. DOI 10.1080/20014091084227.

    Review · Serum GGT is an index of liver dysfunction and a marker of alcohol intake, raised by obstructive liver disease, high alcohol consumption and enzyme-inducing drugs; it breaks down extracellular glutathione, and people with high GGT have higher coronary, diabetes, stroke and mortality risk.

  46. 46.

    Nyblom H, Berggren U, Balldin J, Olsson R. High AST/ALT ratio may indicate advanced alcoholic liver disease rather than heavy drinking. Alcohol and Alcoholism. 2004. PMID 15208167. DOI 10.1093/alcalc/agh074.

    Cross-sectional study · In 313 patients admitted for alcohol withdrawal the AST to ALT ratio was 1.0 or less in 64 percent and rarely 2 or more, while in 48 admitted with alcohol-related cirrhosis 69 percent had a ratio of 2 or more; a high ratio suggests advanced liver disease rather than heavy drinking as such.

  47. 47.

    Conigrave KM, Degenhardt LJ, Whitfield JB, Saunders JB, Helander A, Tabakoff B; WHO/ISBRA Study Group. CDT, GGT, and AST as markers of alcohol use: the WHO/ISBRA collaborative project. Alcoholism: Clinical and Experimental Research. 2002. PMID 11923585.

    Cross-sectional study · In 1,863 people from five countries, carbohydrate-deficient transferrin and GGT performed comparably as markers of alcohol use, AST slightly less well; all detected high-risk better than intermediate drinking; CDT and GGT were influenced by body mass index, sex, age and smoking, and being relatively independent their combination should outperform either alone.

  48. 48.

    Ioannou GN, Dominitz JA, Weiss NS, Heagerty PJ, Kowdley KV. The effect of alcohol consumption on the prevalence of iron overload, iron deficiency, and iron deficiency anemia. Gastroenterology. 2004. PMID 15131790. DOI 10.1053/j.gastro.2004.01.020.

    Cross-sectional study · In NHANES III adults (8,839 non-drinkers; 915 drinking over 2 drinks a day), more than 2 drinks a day carried a significantly higher adjusted prevalence of every iron-overload marker (transferrin saturation over 45 percent, ferritin over 300 ng/mL and combinations), while any alcohol was associated with 40 percent lower risk of iron-deficiency anemia.

  49. 49.

    Choi HK, Atkinson K, Karlson EW, Willett W, Curhan G. Alcohol intake and risk of incident gout in men: a prospective study. Lancet. 2004. PMID 15094272. DOI 10.1016/S0140-6736(04)16000-5.

    Prospective cohort · Among 47,150 men followed 12 years (730 confirmed gout), RR 1.32 at 10 to 14.9 g a day, 1.49 at 15 to 29.9, 1.96 at 30 to 49.9 and 2.53 at 50 or more; beer carried RR 1.49 per daily 12 oz serving, spirits 1.15 per drink, wine 1.04 per 4 oz glass (not significant).

  50. 50.

    Helander A, Péter O, Zheng Y. Monitoring of the alcohol biomarkers PEth, CDT and EtG/EtS in an outpatient treatment setting. Alcohol and Alcoholism. 2012. PMID 22691387. DOI 10.1093/alcalc/ags065.

    Prospective cohort · Among 29 men and 11 women in outpatient treatment, whole-blood phosphatidylethanol was the most sensitive marker of current regular drinking; relapses were detected by PEth alone in 43 percent of cases, by PEth and CDT in 38 percent, and the rest by urinary EtG/EtS.

  51. 51.

    Helander A, Hermansson U, Beck O. Dose-Response Characteristics of the Alcohol Biomarker Phosphatidylethanol (PEth)-A Study of Outpatients in Treatment for Reduced Drinking. Alcohol and Alcoholism. 2019. PMID 31529064. DOI 10.1093/alcalc/agz064.

    Prospective cohort · In 36 outpatients giving drinking diaries and blood every 3 to 4 weeks, PEth 16:0/18:1 correlated best with the past 2 weeks' intake; a change of about 20 g a day moved it by about 0.10 micromol/L, with large scatter between people, so a single value supports only an approximate estimate.

Related

Keep reading.

Reading a number well is the start of the work, not the end of it. If you would like a physician to read yours over time, Explore working with Dr. Tagge.