Alcohol after 40: what the evidence now says

The short answer: the evidence has moved. The long-standing finding that moderate drinkers outlive abstainers has weakened substantially — much of it traced to a study design problem that put people who had quit drinking, often for health reasons, into the non-drinker comparison group. Analyses that correct for this find little or no protective effect, and a cancer risk that rises with dose starting from low intakes. None of that makes a glass of wine a moral failing. It means the health column has less in it than it once appeared to, and the decision is now a clearer trade than it used to be.

What it is

A standard drink is a unit of pure ethanol, and the definitions differ by country in ways that make guidelines hard to compare.

• United States: one standard drink = 14 g of pure alcohol. That’s roughly 355 ml (12 oz) of 5% beer, 150 ml (5 oz) of 12% wine, or 44 ml (1.5 oz) of 40% spirits.
• United Kingdom: one unit = 8 g. A 175 ml glass of 13% wine is about 2.3 units; a pint of 4% beer is about 2.3 units.

Current guidance, for context rather than instruction: US Dietary Guidelines suggest no more than 2 drinks a day for men and 1 for women, on days when alcohol is consumed. The UK advises no more than 14 units a week for everyone, spread over three or more days. Canada’s 2023 guidance moved considerably lower and framed risk as a continuum rather than a threshold — which is closer to where the evidence sits now.

A useful thing to notice: a generous pour at home is often 250 ml, which is roughly two US standard drinks. Most people who track their intake honestly for a week find the number is higher than they estimated.

What the evidence shows, graded

Has been shown to:

• Increase cancer risk in a dose-dependent way. Alcohol is classified as a Group 1 carcinogen — the same category as tobacco and asbestos, which indicates the strength of the evidence that it causes cancer, not that the magnitude of risk is equivalent. The associated cancers are breast, colorectal, liver, oesophageal, oral cavity, pharynx, and larynx. The relationship has no identified safe threshold; risk for breast and oesophageal cancer is detectable at around one drink a day. The mechanism is understood: ethanol metabolises to acetaldehyde, which damages DNA directly.
• Disrupt sleep architecture. Alcohol shortens the time to fall asleep and then degrades the rest of the night — suppressing REM early, fragmenting the second half as it clears, and worsening snoring and breathing disruption. This is the effect most people can verify themselves within a week of changing it.
• Raise blood pressure at higher intakes, and contribute to atrial fibrillation. The link between alcohol and AF episodes is consistent enough that reducing intake is a standard part of managing it.
• Interact with a wide range of medications — covered in its own section below, because after 40 this becomes the most practically relevant item on the list.
• Produce higher blood alcohol per drink as you age. Total body water falls with age, so the same drink is distributed through a smaller volume. Liver metabolism slows somewhat too. The same two glasses that were unremarkable at 35 produce a higher peak concentration and clear more slowly at 60. People usually notice this as increased next-day cost rather than as feeling drunker.

Research suggests:

• The J-curve was largely an artefact. For decades, cohort studies showed light-to-moderate drinkers with lower all-cause and cardiovascular mortality than abstainers. Two problems have since been well characterised. First, abstainer bias — the “never drink” group frequently included people who had stopped because of illness, medication, or a previous drinking problem, making the reference group unusually sick. Second, moderate drinkers in wealthy countries tend to have higher incomes, better diets, and more exercise, and statistical adjustment handles that imperfectly. Meta-analyses restricted to studies with cleaner comparison groups find the protective association shrinks toward nothing.
• Genetic studies point the same way. Mendelian randomisation — which uses inherited variants affecting alcohol metabolism as a natural experiment, and is therefore largely immune to lifestyle confounding — generally finds no cardioprotective threshold, and a rising relationship between alcohol and blood pressure and stroke risk from low intakes upward.
• Heavier drinking is associated with cognitive decline and brain volume loss. Imaging studies find dose-related differences in brain volume extending into what is usually called moderate drinking. Whether this translates into clinically meaningful cognitive change at those levels is not settled.
• Alcohol contributes calories and affects body composition. Roughly 7 kcal/g, no satiety value, and it acutely blunts muscle protein synthesis after training. A regular three-drink evening is a meaningful part of a day’s intake that most people don’t account for.

May help:

• The residual cardiovascular signal. Some analyses still show a small association between light drinking and lower coronary heart disease risk, possibly through effects on HDL cholesterol and clotting factors. It is inconsistent across populations, it is absent in genetic studies, and it does not survive as an overall mortality benefit. If there is a real protective effect, it is small, narrow, and outweighed by the cancer relationship at the same doses.
• Social and psychological benefits. Rarely measured, frequently real. The research apparatus is not built to weigh a shared dinner against a fractional change in relative risk. That weighing is yours to do, and it is a legitimate input rather than a rationalisation.

Medications: the thing that actually changes after 40

Prescription use climbs steadily through midlife, and alcohol interacts with a lot of what gets prescribed. Worth checking specifically:

• Paracetamol / acetaminophen — regular drinking raises the risk of liver injury, including at doses that would otherwise be safe.
• NSAIDs (ibuprofen, naproxen, aspirin) — combined with alcohol, increased risk of gastrointestinal bleeding.
• Sedatives and sleep medication — benzodiazepines, Z-drugs, and older antihistamines all compound alcohol’s sedation and respiratory depression.
• Opioids — the combination is a leading cause of accidental overdose.
• Blood pressure medication — alcohol can amplify the drop and cause dizziness or falls, particularly on standing.
• Anticoagulants — bleeding risk, and heavy intake affects how warfarin is metabolised.
• Antidepressants — varied effects, including reduced efficacy and increased sedation; some older classes have serious interactions.
• Metformin and insulin — alcohol affects glucose regulation and can mask or worsen hypoglycaemia.

None of this means one drink is dangerous on any of these. It means the answer is specific to your list, and a pharmacist can give it to you in about three minutes.

Who it may suit

There is no group for whom the evidence now supports starting to drink for health reasons. That recommendation has been withdrawn essentially everywhere.

For people who already drink and enjoy it, the practical question is not whether to be perfect but where on the dose curve to sit. Risk rises with total volume over time, so the things that move the number most are: how many days a week rather than how much on the occasions you do drink, the size of the pour, and whether heavy sessions occur. Alcohol-free days are the simplest lever most people have, and they also break the habit loop that quietly escalates volume.

Some people find they only notice the cost once it’s gone — better sleep, steadier mood, easier mornings. A four-week experiment is more informative than any risk estimate, and it commits you to nothing.

Who should ask a clinician first

• Anyone with liver disease, pancreatitis, or a history of either.
• Anyone with atrial fibrillation, poorly controlled blood pressure, or heart failure.
• Anyone taking the medications listed above, or several medications generally.
• Anyone pregnant or trying to conceive.
• Anyone with a personal or family history of breast cancer, where the dose-response relationship is one of the better characterised.
• Anyone who has been drinking heavily every day. This one matters medically, not morally: abrupt cessation after sustained heavy daily drinking can cause withdrawal that is genuinely dangerous, including seizures. That is a supervised process, not a willpower problem.

What the marketing overstates

• “Red wine is good for your heart.” Resveratrol is present in quantities far below anything shown to do something in humans — you’d need an implausible volume of wine to approach the doses used in laboratory studies. The wine industry funded a substantial share of the research that built this idea.
• The French Paradox. Largely explained by differences in diet, physical activity, and how causes of death were recorded, rather than by wine.
• “A drink a day keeps the doctor away.” This was a defensible reading of the evidence in 1995. It is not a defensible reading now.
• Low-carb, organic, or “clean” alcohol. The active ingredient is ethanol. Carbohydrate content, sulphites, and farming method don’t change the cancer or cardiovascular relationship.
• In the other direction: headlines announcing that one drink takes a specific number of minutes off your life are also overstating things. These come from modelling exercises with wide uncertainty, and they present relative risk in a format designed to alarm. The honest summary is duller — risk rises with dose, the increases at low intakes are small in absolute terms, and there is no threshold below which it is zero.

If cutting back is harder than expected

Worth stating without commentary: if you have tried to drink less and found it more difficult than you anticipated, that is common and it is treatable. It does not require hitting a crisis, identifying as an alcoholic, or attending anything you don’t want to attend. Effective options include medications that reduce craving, structured brief counselling, and telehealth programmes — most people don’t know the medications exist.

Signs worth taking seriously: drinking more than you intended once you start, needing more for the same effect, feeling shaky, sweaty, anxious, or nauseated the morning after or when you go without, or repeatedly deciding to cut down and not managing it. The NIAAA maintains a free, non-commercial directory of evidence-based treatment options if you want to look privately.

When to talk to a clinician

Bring it up if you take regular medication and drink at all — the interaction question is worth two minutes and the answer is specific to you.

Also worth raising: liver enzymes flagged on a routine blood test, new or worsening reflux, high blood pressure that isn’t responding to treatment, an irregular heartbeat, worsening sleep or new snoring, unexplained bruising, or falls and unsteadiness.

Seek advice before stopping, not after, if you have been drinking heavily every day — withdrawal from that pattern can require medical supervision.

And if you want a straight answer about your own numbers rather than population averages, tell your clinician what you actually drink in a typical week. The usual assumption is that the figure given is an underestimate, so the accurate one gets you better advice.