Why I no longer tell patients one or two drinks is fine
The evidence that light drinking is protective has mostly fallen apart. Two large 2022 and 2023 analyses show measurable harm at levels most people consider moderate.
For twenty years I have told patients that one or two drinks a day was, as far as the evidence went, fine. For most of that time, I believed it. The research seemed to support it. Moderate drinkers lived longer than teetotal drinkers in population studies. Red wine was said to protect the heart. I even occasionally used that line with patients who were stressed about having a glass of wine with dinner.
I don’t say that anymore.
Not because I’ve become puritanical about alcohol. Because the research has genuinely changed, and honest advice needs to change with it.
Why the old story held for so long
The “light drinking is protective” narrative came out of observational studies that compared people who drank moderately with people who didn’t drink at all. The moderate drinkers had lower rates of cardiovascular disease and sometimes lower all-cause mortality. Red wine was singled out because of polyphenols like resveratrol.
The story was biologically plausible enough that it survived for decades. But it rested on a quiet methodological flaw. The “teetotal” reference group included two very different populations. Lifelong non-drinkers, who are often healthier than average, and ex-drinkers, who quit because they were already sick.
When you don’t separate those two groups, the ex-drinkers drag the non-drinker category toward worse outcomes, and moderate drinkers look protected by comparison. When you do separate them, the apparent benefit of light drinking largely disappears, or becomes very small compared with the harms on the other side of the ledger.
Better-designed more recent studies have done that separation. And the picture has shifted.
What the 2023 cancer meta-analysis shows
A systematic review and meta-analysis published in 2023 pooled 139 prospective cohort studies on alcohol and cancer. The question it asked was simple. Is there a dose of alcohol below which cancer risk isn’t raised?1
The answer the data gave: no.
Cancer risk rises continuously from low-level drinking upward. There is no plateau at the bottom. The cancers most consistently linked to alcohol intake are:
- Breast cancer (one of the biggest public-health signals)
- Oesophageal cancer
- Oral cavity and pharyngeal cancer
- Liver cancer
- Colorectal cancer
The mechanism is reasonably well-understood. When your body metabolises alcohol, the first step produces acetaldehyde. Acetaldehyde binds directly to DNA, impairs DNA repair, and is classified as a Group 1 carcinogen by the International Agency for Research on Cancer. The body clears it quickly, but every drink generates more.
This isn’t a fringe concern. Alcohol is a known carcinogen, in the same regulatory category as tobacco smoke and asbestos.
What the 2022 brain imaging data shows
A 2022 analysis of roughly 36,000 adults in the UK Biobank examined brain MRI scans against self-reported alcohol intake. The finding, put bluntly: grey and white matter volume reductions begin at around one drink per day and scale roughly linearly as intake increases.2
The effect sizes at one drink are small. At three or more drinks per day they are substantial. But the curve is continuous from the low end. There’s no flat zone of safety at the bottom.
Those structural changes matter for function over time. The regions most affected include parts of the frontal cortex involved in decision-making, impulse control, and executive function, and subcortical structures involved in memory. The research on cognitive decline trails the structural imaging but consistently points the same direction.
The practical dose-response
The dose-response matters because it tells you what reducing (rather than eliminating) actually buys.
Reducing from three or more drinks a day to one a day produces a real drop in expected harm. The effect sizes at three or more drinks a day are much larger than at one. The curve is steeper at the top than at the bottom.
Reducing from one drink a day to two a week produces a smaller but still real drop.
Eliminating entirely removes the dose-dependent risk from both the brain imaging and the cancer data.
None of those moves are all-or-nothing. The people I see in clinic who successfully cut down rarely stop abruptly. They introduce alcohol-free days, reduce the size of pours, and stop stacking alcohol with other cardiovascular stressors like poor sleep or sauna.
What I actually say in clinic now
The conversation I have with patients now has three parts.
First, the facts. Alcohol is a known carcinogen. Low doses affect the brain. The older protective-effect story for the heart has largely fallen apart under better methods.
Second, the patient’s choice. I don’t tell people to stop drinking. Most of my patients drink socially, enjoy it, and that is information they get to weigh against the risks. My job is to make sure they’re weighing accurate risks, not the 1990s version.
Third, the practical moves. Two alcohol-free days a week. Keep pours honest. Don’t stack alcohol with other acute cardiovascular stressors. Reduce, don’t necessarily eliminate. Reducing matters more than it used to, and eliminating matters more than it used to.
The bottom line
The old message was “light drinking is fine, possibly good for the heart.” That message was based on flawed reference groups and it hasn’t survived better data.
The honest updated version is this. Alcohol is a dose-dependent carcinogen with measurable effects on the brain starting at roughly one drink a day. Whether and how much you drink is up to you. But the risks start lower than we used to say, and reducing is worth doing even if you don’t plan to stop.
That is what I tell myself. It is what I now tell my patients.
Frequently asked questions
Is there a safe level of alcohol?
For cancer risk, no. A 2023 systematic review and meta-analysis of 139 prospective cohort studies found no threshold below which cancer risk isn't elevated. Risk rises continuously from low-level drinking upward. For cardiovascular disease, the older 'one or two drinks is protective' story has been substantially weakened by methods that control for ex-drinkers (people who stopped because they were unwell). Once that confounding is removed, the protective effect largely disappears or becomes very small.
Does one drink a day actually shrink your brain?
In a 2022 UK Biobank analysis of 36,000 adults, grey and white matter volume reductions were measurable starting at around one drink per day and scaled roughly linearly with increasing intake. The relationship was not flat-then-linear. There was no safe zone at the low end. The effect sizes at one drink were small but real. At three or more drinks per day, the structural changes were substantial.
What about the French Paradox and red wine for heart health?
The French Paradox is a correlation that survived mostly because the original data didn't separate light drinkers from people who had stopped drinking for health reasons. When you properly exclude ex-drinkers from the reference group, the apparent heart benefit of light drinking largely disappears. The polyphenols in red wine (resveratrol and others) have some biological activity in cell studies, but the doses required for a meaningful effect are far higher than you'd get from drinking wine.
How does alcohol cause cancer?
The main mechanism is acetaldehyde, the first metabolite your body produces when it breaks down alcohol. Acetaldehyde binds directly to DNA and interferes with DNA repair. It is classified as a Group 1 carcinogen by the International Agency for Research on Cancer. The body clears it quickly, but every drink generates more. Cancers most strongly linked to alcohol are breast, liver, mouth, throat, oesophagus, and colorectal. The dose-response is continuous. There is no threshold below which cancer risk isn't raised.
Is it worse to drink every day or binge drink at weekends?
Both carry risk, through partly different mechanisms. Daily drinking, even at low levels, produces the cumulative cancer and brain-volume effects described above. Binge drinking produces acute cardiovascular strain, higher injury risk, and a different pattern of liver and brain harm. For total health risk, consistent daily drinking and weekend binge drinking at equivalent weekly totals both elevate risk, though the specific diseases each raises are weighted differently.
What do I tell patients who don't want to stop drinking entirely?
The same thing I tell myself: the risk starts lower than we used to say, and reducing is better than not reducing. Practical targets that make a real difference: two or more alcohol-free days a week, drink-free weekdays, capping weekly intake well below old 'safe' guidelines, and never stacking alcohol with poor sleep, medications, or hot environments like sauna. Reducing matters more than eliminating, and eliminating matters more than it used to.
References
- 1.
Cancer risk based on alcohol consumption levels: a comprehensive systematic review and meta-analysis · Jun S, Park H, Kim UJ, et al. · Epidemiology and Health (2023) PubMed PMID 37905315
- 2.
Associations between alcohol consumption and gray and white matter volumes in the UK Biobank · Daviet R, Aydogan G, Jagannathan K, et al. · Nature Communications (2022) PubMed PMID 35246521