How much alcohol is actually safe to drink?
Sharing a bottle of wine a night exceeds current public-health guidance. The biology behind that is mainstream and solid; the argument that the social benefit offsets it is his own weighting.
What this episode covers
- Alcohol is toxic to cells, which is why it can damage tissue from the mouth down.
- The harm also rises faster than the amount you drink, so doubling your intake more than doubles the risk.
- The amount that feels normal is often well past the guidelines.
Confidence in this episode
Everything about how much to believe this episode, in one place.
The core biology is mainstream and well supported; the headline ranking is an expert panel's weighting rather than a measurement, and the speaker sells an alcohol substitute.
- Alcohol is directly toxic to cells. It is used to kill bacteria on skin, and the same property damages tissue from the mouth to the liver. This is not disputed.
- Harm rises faster than dose. He puts 14 to 28 units a week at roughly triple the harm, and 28 to 56 at more than quadruple (his figures).
- Alcohol raises blood pressure, and cutting down is one of the more tractable ways to bring it back down. This is standard clinical advice, not his own position.
- Cutting alcohol tends to lower cholesterol as well as blood pressure. He is unambiguous, and this is mainstream.
- Alcohol disrupts sleep architecture even at low doses. It speeds sleep onset, pulls deep sleep early, and shortens the night through a mild overnight withdrawal.
- Whether moderate drinking is a net gain. He argues the social benefit can outweigh small physical harms; the WHO position is that no level is safe. This is a values judgement as much as an evidence one.
- The 'most harmful drug' ranking. It aggregates 16 kinds of harm including harm to other people, and the weighting is an expert panel's judgement, not a measurement.
- Where the line actually sits. He says the odd binge may be worse than the same amount spread out, and declines to name a threshold.
- How much of the brain change recovers if you stop, and over what timescale. He says we don't know.
- His commercial position. He invented and sells Sentia, a herbal alcohol substitute, which is the product a drink-less message points toward.
Why it matters
Most people picture alcohol harm as addiction, or as something that happens to other people. The evidence points somewhere more ordinary: steady drinking inside what feels like a normal social life. If alcohol stiffens arteries, raises blood pressure, inflames the brain, strains the liver and adds to cancer risk, then cutting back may move several numbers at once. The challenge is that the harm builds slowly and depends on your dose, your genes and your pattern of drinking, which most people never track.
What stands out
- Most people with alcohol-related liver damage never thought they had a problem: he cites a hepatologist's decades-long cohort in which a third were middle-class people sharing a bottle of wine a night (as he reports it).
- Doubling your drinking does more than double the harm. The curve bends upward rather than running straight, so the last few units of the week cost more than the first few (his figures).
- The French paradox mostly isn't the wine. He puts any benefit at 100ml a day at most, and suspects the alcohol cancels out the polyphenols anyway (his view, not formally tested).
Best-supported action
The single highest-leverage move from this episode, anchored in the strongest evidence the speaker presents.
His one tip: write down every drink for a week, then cut the ones that gave you nothing in hindsight. He estimates that is about half of what most people drink, and finding out costs you nothing.
Where to start
Small low-friction starters covering the main moves from this episode.
- Notice which drinks you would genuinely miss, and which ones you are just holding.
- Have the wine after the meal rather than with it, so you actually taste it.
- If someone close to you has ever commented on your drinking, sit with that rather than defending it.
Other supported actions
Further actions discussed in this episode, ordered from strongest to weakest evidence. This is one expert's view, the full topic compares and ranks across experts.
- His baseline is the current UK guidance rather than a number of his own: under 14 units a week, spread out, with at least two alcohol-free days. He adds that the pattern matters as much as the total, since one heavy night may do more damage than the same amount spread across the week.Moderate evidence
- Where blood pressure or cholesterol is already high, he suggests cutting alcohol before adding anything else, then re-testing after a few weeks. He calls it the most tractable lever available, though how far your numbers move depends on how much you were drinking.Moderate evidence
- At more than a bottle of wine a day, he suggests stepping down to half a bottle for a few days rather than stopping outright. Withdrawal from that level can be dangerous, so this is one to take to your doctor before you start.Moderate evidence
Full context, impact ratings, and timing — available in related topics
Questions to take to your doctor
- Given my blood pressure, would cutting down on alcohol for a few weeks be worth trying before we add another medication?
- Given how much I drink, is it safe for me to just stop, or should I step down first?
- Given my family history of breast cancer, does my drinking change my risk enough to act on?
Full doctor prep with ranked questions available in the full topic page
Context
Academic psychiatrist and neuropsychopharmacologist who has spent four decades on how drugs act on the brain, and who treats alcohol as a drug like any other rather than as a social category exempt from that question. Strongest on the pharmacology and the dose-response, where he sits squarely in the mainstream; his distinctive move is to weigh social benefit against physical harm explicitly rather than leave it out, which takes him past what the evidence alone settles, and he has a commercial interest in a herbal alcohol substitute he invented.
This episode does not establish where the safe line sits. He declines to name one, and says the pattern may matter as much as the total: the same weekly amount taken as one heavy night may do more damage than spread across a week. The brain-volume finding he cites comes from a long-running observational study of civil servants, which can show that heavier drinkers have smaller brains but cannot show the drinking caused it. Individual response varies a lot, and tolerance hides it. He is explicit that feeling fine is not evidence of being fine.
The headline that alcohol is the most harmful drug is a ranking, not a measurement. It comes from an expert panel scoring 16 kinds of harm, nine to the user and seven to other people, then adding them together. Alcohol tops the list partly because so many people drink, which loads the harm-to-others side. The exercise has been repeated in five countries with the same answer, but the weighting is a judgement, and someone weighting personal harm above social harm would land somewhere different.
He invented and sells Sentia, a herbal drink marketed as an alcohol alternative, and mentions drinking it himself during the week. This does not invalidate decades of published work on alcohol harm, and the two are worth judging separately. It is still worth knowing that the person making the case for drinking less also sells the substitute.
None of this means you should change or stop your current treatment on your own. If you drink heavily, stopping suddenly can be dangerous, and that is a conversation for your doctor rather than a decision to take from a podcast.
Overall evidence profile: the toxicity and the dose-response are mainstream and solid, resting on large observational data rather than trials. Where he goes beyond the record is the weighting, not the biology.
Where people go wrong
- Treating a shared bottle of wine a night as normal because everyone around you does it.That is roughly two and a half times the weekly guideline. In the cohort he cites, a third of middle-class patients with alcohol-related cirrhosis had never considered themselves drinkers with a problem.
- Going back to your old amount after a dry month.You lose tolerance while you are off it, so the old amount hits far harder. He describes fatal overdoses that follow exactly this pattern.
What to expect over time
- First few nightsSleep may be worse before it is better. He describes a couple of unsettled nights and some anxiety.
- Within a week or twoMost people settle. He describes sleep and morning energy recovering over about a week where there is no dependence.
- After a monthBlood pressure and cholesterol may drift down. He says most people do not go back to their old amount afterwards.