Red wine owes its color to the grape skins. Dark grapes carry anthocyanin pigments in the skin, and red winemaking extracts color and flavor from the skins during fermentation. That skin contact is also why red wine has much more resveratrol than white wine, which is fermented after the skins are removed. A 5 fl oz glass (147 g) holds about 28 mg of anthocyanins and 51 mg of total flavonoids, but 88% of its 125 calories come from alcohol.
Two stories made red wine famous. The first is the French paradox: in 1992 a Lancet paper noted that France had a diet high in saturated fat but low heart disease deaths, and suggested wine might explain part of it. A 1999 BMJ analysis argued that a time lag in diet change and under-recorded heart disease explained much of the gap, with wine playing little part. The second is resveratrol, a compound from grape skins that extended yeast lifespan by 70% in a 2003 lab study.
A glass holds a tiny dose of resveratrol, about 0.3 mg on average, while the supplement trials used 75 mg or more a day, and the human results have been mostly null. Alcohol is a Group 1 carcinogen, the World Health Organization says no level of alcohol is safe, and low intake already raises breast cancer risk. This page shows the evidence on both sides and does not recommend drinking.
Health benefits of red wine
The best-supported benefit finding is small: in a 2-year trial in people with type 2 diabetes who did not drink, red wine with dinner raised HDL cholesterol a little. The best-supported harm finding is stronger: alcohol raises cancer risk, including breast cancer at low intake, and newer analyses that correct for sick-quitter bias find no clear mortality benefit for light drinking. None of the studies below reaches strong evidence. Here is what 30 studies found, strongest evidence first, from the CASCADE trial and resveratrol supplements to cancer, mortality and the French paradox. None of it is advice to start or keep drinking.
Drinking 150 mL of red wine with dinner for 2 years raised HDL cholesterol by 0.05 mmol/L in people with type 2 diabetes who did not drink before
The CASCADE trial randomly assigned 224 alcohol-abstaining adults with well-controlled type 2 diabetes to mineral water, white wine or red wine with dinner, all on a Mediterranean diet. Red wine also lowered the total cholesterol to HDL ratio by 0.27, while blood pressure, liver function and body fat did not differ. Participants knew what they drank, and it measured cholesterol and blood sugar, not heart attacks or cancer.
In trials, drinking alcohol raised HDL cholesterol by 0.094 mmol/L and lowered fibrinogen by 0.20 g/L
Meta-analysis of interventional studies of 13 heart disease markers in adults without known cardiovascular disease (44 of 63 eligible studies were pooled). Adiponectin also rose and triglycerides did not change, with similar results for wine, beer and spirits. The authors call this indirect support for protection from moderate drinking, and it measured blood markers only, not heart attacks or cancer.
Cutting alcohol intake by about half lowered systolic blood pressure by 5.5 mm Hg in people who drank six or more drinks a day
Meta-analysis of 36 trials with 2865 people, mostly men (2464). In people who drank two or fewer drinks a day, cutting back did not change blood pressure significantly, and diastolic pressure fell 3.97 mm Hg in the heaviest drinkers. It covers alcohol in general, not wine alone.
An umbrella review concluded that current evidence does not support resveratrol supplements for diabetes, metabolic syndrome or fatty liver
Review of meta-analyses of randomized trials covering 1476 people with type 2 diabetes, 727 with metabolic syndrome and 271 with fatty liver. For almost all outcomes the effect was trivial or the certainty low. The one clearer result was a 1.05-point fall in HbA1c within 12 weeks (6 trials), which the authors note comes from short, small studies. These were resveratrol supplements, not wine.
Show 26 more studies
Resveratrol supplements did not lower CRP, cholesterol or blood pressure in a pooled analysis of trials
Meta-analysis of 10 randomized trials (11 treatment arms). HDL cholesterol was slightly lower on resveratrol, by 4.18 mg/dL, and the authors call for larger, well-designed trials. These were supplements, not wine.
Resveratrol supplements lowered fasting blood glucose by 0.29 mmol/L in people with type 2 diabetes, but HbA1c barely changed
Meta-analysis of 9 randomized trials with 283 people. Insulin fell 0.64 U/mL and blood pressure fell too, while changes in HbA1c and in LDL and HDL cholesterol were negligible. The glucose effect was larger at doses of 100 mg a day or more. These were supplements, not wine.
Across 17 trials, resveratrol supplements did not significantly change blood pressure, except at 300 mg a day or more
Meta-analysis of randomized trials. Systolic pressure changed by -2.5 mm Hg overall, which was not significant, and results varied a lot between trials. Significant drops appeared only in subsets taking 300 mg a day or more and in people with diabetes. These were supplements, not wine.
Dealcoholized red wine lowered systolic and diastolic blood pressure in men at high heart risk
Randomized crossover trial in 67 men at high cardiovascular risk, who each had red wine (30 g of alcohol a day), the same amount of dealcoholized red wine and gin (30 g of alcohol a day) for 4 weeks at a time. The paper reports the pressure drop after the dealcoholized wine and links it to higher plasma nitric oxide. It was a short report in men only, and it measured blood pressure, not heart attacks.
Red wine and dealcoholized red wine both lowered insulin resistance, while HDL cholesterol rose after red wine and gin
Same crossover trial as above, in 67 men at high cardiovascular risk, with 30 g of alcohol a day in the red wine and gin periods. Fasting glucose did not change and lipoprotein(a) fell after red wine. The authors credit the non-alcoholic part of red wine, mainly polyphenols, for the insulin effect.
12 weeks of resveratrol supplements at 75 mg a day did not improve metabolic function in healthy postmenopausal women
Double-blind, placebo-controlled trial in nonobese postmenopausal women with normal glucose tolerance. Resveratrol did not change body composition, resting metabolic rate, blood fats or inflammation markers, and it did not raise insulin sensitivity in liver, muscle or fat. It was a supplement, not wine, in women with healthy metabolism.
After correcting for bias from former drinkers, low-volume drinkers had no significant mortality advantage over lifetime nondrinkers
Meta-analysis of 107 cohort studies with about 4.8 million people and 425,564 deaths. Drinkers of up to 24 g of alcohol a day had a relative risk of 0.93 (not significant), while risk rose at 45 to 64 g a day (1.19) and at 65 g or more (1.35). Female drinkers had a 22% higher risk than female lifetime nondrinkers.
Low-volume drinkers seemed to live longer, but after correcting for abstainer bias the benefit disappeared
Meta-analysis of 87 cohort studies with 3,998,626 people and 367,103 deaths. The raw result looked like the classic J-curve (relative risk 0.86 at 1.3 to 24.9 g a day), but after adjustment it was 0.97. Former drinkers, who often quit because of ill health, had a higher risk (1.22) and had been counted among the abstainers.
People with a gene variant linked to less drinking had 10% lower odds of coronary heart disease, even among light to moderate drinkers
Mendelian randomization of 56 studies with 261,991 people, which uses inherited differences in drinking to avoid the usual biases. Carriers of the ADH1B A-allele drank 17.2% fewer units a week and had lower systolic pressure (0.88 mm Hg) and lower odds of ischemic stroke (0.83). The authors conclude that cutting down is beneficial even for light to moderate drinkers.
In China, genes that predict heavier drinking raised stroke risk steadily, with no protective dip for light drinkers
Prospective study of 512,715 adults, 161,498 of them genotyped. In the usual analysis men who drank about 100 g of alcohol a week had lower stroke and heart attack risk, but the genetic analysis showed a steady rise in ischemic stroke risk (relative risk 1.27 per 280 g a week) and intracerebral hemorrhage (1.58), and little net effect on heart attack. Most drinking was spirits, not wine.
Among drinkers, the lowest risk of death was at about 100 g of alcohol a week or less, and stroke risk rose with every extra drink
Pooled data on 599,912 current drinkers in 83 studies. Each extra 100 g a week went with a 14% higher stroke risk (hazard ratio 1.14), while heart attack risk was lower. Drinking more than 100 g a week went with a shorter life expectancy at age 40, by about 6 months at 100 to 200 g and 4 to 5 years above 350 g a week.
Women who drank 3 to 6 drinks a week had a 15% higher risk of breast cancer
Nurses’ Health Study of 105,986 women, with 7690 invasive breast cancers diagnosed from 1980 to 2008. The rise was statistically significant at 5.0 to 9.9 g of alcohol a day (relative risk 1.15), and drinking earlier and later in adult life each added risk. Binge drinking mattered beyond total intake.
Light drinking, up to 1 drink a day, was linked to a 5% higher risk of female breast cancer and with higher risks of mouth, throat and esophageal cancer
Meta-analysis of 222 articles. Relative risks were 1.05 for breast cancer, 1.17 for cancer of the oral cavity and pharynx and 1.30 for esophageal squamous cell carcinoma, with no link for colorectal, liver or larynx cancer. The authors estimate about 5000 breast cancer deaths worldwide in 2004 were attributable to light drinking.
In more than a million British women, each extra drink a day was linked to 12% more breast cancer, and the trend was similar in women who drank only wine
Million Women Study of 1,280,296 women followed for an average of 7.2 years, with 68,775 invasive cancers. Total cancer rose 6% per 10 g of alcohol a day, and the authors estimate about 15 extra cancers per 1000 women up to age 75 for each extra daily drink, 11 of them breast cancers.
An estimated 741,300 new cancers worldwide in 2020 were caused by alcohol, including 103,100 from moderate drinking
Modeling study using relative risks from World Cancer Research Fund reviews and 2020 cancer data. Alcohol accounted for 4.1% of all new cases, with the most cases in the esophagus, liver and breast (98,300). Moderate drinking meant under 20 g a day, and drinking up to 10 g a day alone contributed 41,300 cases.
People who drank 14 to 21 units a week for 30 years had three times the odds of hippocampal atrophy as abstainers
Whitehall II cohort of 550 men and women followed for 30 years, with brain scans at the end. Moderate drinkers had odds of 3.4 for right-sided hippocampal atrophy, and light drinking (1 to under 7 units a week) showed no protective effect over abstinence. Higher drinking also went with faster decline in verbal fluency.
In the PREDIMED trial, a Mediterranean diet that included wine went with a lower risk of death, but the link weakened in analyses built to avoid abstainer bias
Observational analysis of 7447 high-risk people in PREDIMED and 23,133 people in the Spanish SUN cohort. Good diet adherence including wine had a hazard ratio of 0.67 for death, against 0.77 without wine. The lower risk was absent in people who drank three or more glasses a day, the SUN cohort showed no significant link with cardiovascular disease, and the authors advise caution.
PREDIMED participants whose urine tartaric acid reflected 12 to 35 glasses of wine a month had half the rate of cardiovascular events
Case-cohort study of 1232 high-risk participants, using urinary tartaric acid, an objective marker of wine intake, instead of questionnaires. The rate was also lower at 3 to 12 glasses a month (hazard ratio 0.62), and links with self-reported wine intake were less significant. It shows a link, not proof, and wine drinkers may differ in other habits.
Wine drinking showed a J-shaped link with vascular risk, with the lowest risk, 31% lower, at 21 g of alcohol a day
Meta-analysis of 16 studies of wine. A similar J-shaped curve appeared for beer (13 studies) but not for spirits (10 studies). This is the classic J-curve finding that the bias-corrected analyses above question.
In older adults in the Chianti wine region, urinary resveratrol levels did not predict who died over 9 years
InCHIANTI cohort of 783 people aged 65 or older, of whom 268 died. Death rates from the lowest to the highest quartile of resveratrol metabolites were 34.4%, 31.6%, 33.5% and 37.4%. Resveratrol was not linked to inflammation markers, heart disease or cancer, and the authors say amounts from a Western diet did not have a substantial influence.
The paper behind the French paradox suggested that wine may explain part of France’s low heart disease despite a diet high in saturated fat
A hypothesis paper, not a trial. It argued that alcohol intake of 20 to 30 g a day, as in France, could cut coronary heart disease risk by at least 40%, probably through effects on platelets, and noted that HDL cholesterol was no higher in France.
A 1999 analysis found that a time lag in diet and under-recorded heart disease explained much of the French paradox, and wine little
Analysis of national statistics comparing France with Britain and other countries. The authors judged that alcohol, and red wine in particular, explained little of the gap, and expected French heart disease deaths to rise as the lag ran out. The PubMed record has no abstract, so this summary follows the paper itself.
Red wine contains an average of 1.9 mg of trans-resveratrol per liter, and up to 14.3 mg per liter at the highest level reported
Review of published chemical analyses of red wine, not a study in people. At the average, a 5 fl oz glass (about 148 mL) holds roughly 0.3 mg, and about 2 mg at the highest level reported. The resveratrol supplement trial above used 75 mg a day.
Resveratrol did not extend the lifespan of mice in the National Institute on Aging testing program
Lifelong treatment of genetically varied mice at three test sites, starting at 4 months of age, with resveratrol and four other agents. None of the five had a significant effect on lifespan in males or females at the doses tested. Green tea extract might have lowered midlife deaths in females only.
Resveratrol improved health and survival of mice fed a high-calorie diet
Study in middle-aged mice, not people. Resveratrol shifted their physiology toward that of mice on a standard diet, with better insulin sensitivity and motor function, and it opposed the diet’s effects in 144 of 153 altered gene pathways. Human supplement trials above have been mixed.
Resveratrol extended yeast lifespan by 70% and activated a sirtuin enzyme in the lab
Lab study in yeast and cells, not people. It reported resveratrol activating the enzyme SIRT1 and mimicking calorie restriction in yeast, which launched the idea that red wine might slow aging. The human trials above are mixed or null.
Who should be careful with red wine
No level of alcohol is considered safe for health by the WHO, and the less you drink, the lower the risk. Do not drink at all, or ask your doctor first, if one of these applies:
- Pregnancy or trying to conceive. The NHS advises no alcohol if you are pregnant or planning to become pregnant, because it can cause lifelong harm to the baby, including fetal alcohol spectrum disorder.
- Under the legal drinking age. The NIAAA notes that the brain keeps developing well into the 20s and that alcohol can interfere with that development.
- Medicines or liver disease. Alcohol can interact with many medicines, including some sold over the counter, so ask a pharmacist. Alcohol-related liver disease does not always cause symptoms in its early stages.
- Alcohol use disorder or trouble cutting down. These studies are no reason to drink. In the US, the NIAAA Alcohol Treatment Navigator helps find care, and in the UK the NHS suggests starting with a GP; stopping suddenly after heavy, dependent drinking can be dangerous, so get advice first.
How to choose, store and eat red wine
- Know the serving: a 5 fl oz glass has 125 calories and 15.6 g of alcohol. At 12% alcohol it is one US standard drink (14 g); a bigger pour or a stronger wine holds more.
- Dealcoholized wine: the blood pressure and insulin results above came from dealcoholized red wine in 67 men, and none of the trials tested the alcohol-free wines sold in shops.
- Sulfites: US wine labels must say so when sulfites reach 10 parts per million, a rule meant for people who are sensitive to them.
Check the alcohol percentage on the label rather than assuming a glass is a standard drink. An opened bottle can spoil in about three days, so keep it corked and finish it soon.
Sources
- What is a standard drink? NIAAA.
- Wine labeling: declaration of sulfites TTB, US Treasury.
- Savor and save wine, one glass at a time Purdue University. 2019.
- No level of alcohol consumption is safe for our health WHO Regional Office for Europe. 2023.
- Drinking alcohol while pregnant NHS.
- Get the facts about underage drinking NIAAA.
- Harmful interactions: mixing alcohol with medicines NIAAA.
- Alcohol-related liver disease (ARLD) NHS.
- Alcohol Treatment Navigator NIAAA.
- Alcohol support NHS.
- Chiva-Blanch G, Urpi-Sarda M, Ros E, et al. Dealcoholized red wine decreases systolic and diastolic blood pressure and increases plasma nitric oxide: short communication. Circ Res. 2012.
History of red wine
Wine is older than writing. Grapevines were domesticated about 11,000 years ago in Western Asia and the Caucasus, according to a 2023 genome study of 3525 cultivated and wild vines. The earliest chemical evidence of grape wine comes from pottery in Georgia, in the South Caucasus, dated to about 6,000 to 5,800 BC. Very large jars from that culture probably served for fermenting, aging and serving the wine.
Red wine is defined by its skins. Most grape juice is greenish-white, and the red color comes from pigments in the skin, so making red wine means extracting color and flavor from the skins. That is also why red wines carry more of the skin compound resveratrol than whites.
The modern health story began in France. In 1991 Serge Renaud of Bordeaux University presented the idea behind the French paradox, and a 1992 Lancet paper by Renaud and Michel de Lorgeril suggested wine might explain part of France’s low heart disease. In 1999 Malcolm Law and Nicholas Wald argued that a time lag in diet and under-recorded heart disease explained much of the paradox.
Resveratrol was first isolated in 1939 by the Japanese chemist Michio Takaoka, from a plant called white hellebore, not from wine. A 2003 Nature paper that reported resveratrol extending yeast lifespan made it famous. David Sinclair, a co-author, co-founded Sirtris Pharmaceuticals in 2004, which GlaxoSmithKline bought in 2008 for $720 million and shut down in 2013 without a successful drug.
Not all of the resveratrol literature held up. In 2012 the University of Connecticut Health Center announced that its review board found Dipak Das, a researcher who published many resveratrol studies, guilty of 145 counts of fabricating or falsifying data, and more than twenty of his papers have since been retracted. In mice, the National Institute on Aging testing program found no lifespan gain from resveratrol.
Sources
- Dong Y, Duan S, Xia Q, et al. Dual domestications and origin of traits in grapevine evolution. Science. 2023.
- McGovern P, Jalabadze M, Batiuk S, et al. Early Neolithic wine of Georgia in the South Caucasus. Proc Natl Acad Sci U S A. 2017.
- Red wine. Wikipedia.
- French paradox. Wikipedia.
- Renaud S, de Lorgeril M. Wine, alcohol, platelets, and the French paradox for coronary heart disease. Lancet. 1992.
- Law M, Wald N. Why heart disease mortality is low in France: the time lag explanation. BMJ. 1999.
- Howitz KT, Bitterman KJ, Cohen HY, et al. Small molecule activators of sirtuins extend Saccharomyces cerevisiae lifespan. Nature. 2003.
- Resveratrol. Wikipedia.
- Dipak K. Das. Wikipedia.
- Strong R, Miller RA, Astle CM, et al. Evaluation of resveratrol, green tea extract, curcumin, oxaloacetic acid, and medium-chain triglyceride oil on life span of genetically heterogeneous mice. J Gerontol A Biol Sci Med Sci. 2013.
Frequently asked questions about red wine
Is red wine good for you?
Not clearly, and not for everyone. A 5 fl oz glass has 125 calories and 15.6 g of alcohol, which is 88% of the calories, with almost no protein or fiber. In a 2-year trial in people with type 2 diabetes who did not drink, red wine with dinner raised HDL cholesterol by 0.05 mmol/L, but analyses that correct for sick-quitter bias found no significant mortality benefit for low-volume drinkers (relative risk 0.97). If you do not drink, these studies are no reason to start.
Is red wine good for your heart?
The evidence is contested. A meta-analysis found 31% lower vascular risk at 21 g of alcohol a day, but genetic studies point the other way: in China, genetically predicted heavier drinking raised stroke risk steadily, and in a 261,991-person analysis a variant linked to less drinking went with 10% lower odds of coronary heart disease. In PREDIMED, a diet with wine went with lower risk, but the link weakened in analyses built to avoid abstainer bias.
Does red wine cause cancer?
Alcohol does, and the type of drink does not change that. The WHO says any beverage containing alcohol poses a cancer risk, and alcohol is a Group 1 carcinogen. In the Nurses’ Health Study, breast cancer risk was 15% higher at 5.0 to 9.9 g of alcohol a day, and a meta-analysis found 5% higher risk at up to 1 drink a day. In the Million Women Study the trend was similar in women who drank only wine.
Does red wine help you live longer?
Not in the best-corrected analyses. A 2023 meta-analysis of 107 studies found no significant mortality benefit for low-volume drinkers (relative risk 0.93), and risk rose at 45 to 64 g a day (1.19). In older adults in the Chianti wine region, urinary resveratrol did not predict who died, and resveratrol did not extend the lifespan of mice in the National Institute on Aging program.
How much resveratrol is in a glass of red wine?
Very little. Published analyses average 1.9 mg of trans-resveratrol per liter of red wine, so a 5 fl oz glass holds roughly 0.3 mg, and about 2 mg at the highest level reported. The supplement trials used 75 mg a day or more, and the blood pressure meta-analysis found significant drops only at 300 mg a day or more. At the average, 75 mg would take about 270 glasses, a figure calculated from these two numbers.
How many calories are in a glass of red wine?
A 5 fl oz glass (147 g) has 125 calories, or 85 per 100 g. Alcohol supplies 15.6 g of that, 88% of the calories, and there are 3.8 g of carbohydrate (0.9 g of sugars), 0.1 g of protein, no fat and no fiber. Its nutrient density score of 41 is low, and the minerals that stand out are molybdenum (33% of the daily value) and chromium (13%).
Is red wine healthier than white wine?
No study here shows that it is. In the CASCADE trial both wines helped blood sugar in people who clear alcohol slowly, and the trial’s HDL gain was reported for red wine. The WHO says it is the alcohol that causes harm, not the beverage, and the Million Women Study found a similar cancer trend in wine-only drinkers. Red wine does have more skin pigments, about 28 mg of anthocyanins per glass.
Is non-alcoholic red wine healthier?
It removes the alcohol, but the evidence is thin. In a crossover trial of 67 men at high heart risk, dealcoholized red wine lowered blood pressure and insulin resistance, with 4 weeks on each drink. That is small, short and in men only, it measured markers rather than heart attacks, and shop-bought alcohol-free wines were not tested.