Alcohol

TL;DR. The scientific position on alcohol changed substantially in the last decade, and the direction was one way. The apparent heart protection from moderate drinking has largely dissolved once studies stopped counting sick ex-drinkers as abstainers, and genetic studies that avoid that problem find no cardiovascular benefit. Alcohol is a Group 1 carcinogen with no safe threshold for cancer risk, and it causes cancer of the mouth, throat, oesophagus, liver, colon, and breast. It is also woven into human culture everywhere and most people who drink will continue to. This chapter is about knowing what you are choosing.

1. What it is and what it does

Ethanol ($\mathrm{C_2H_5OH}$), produced by yeast fermenting sugar. It is small, water-soluble, lipid-soluble, and uncharged, which is why it crosses every membrane in your body including the blood-brain barrier and the placenta within minutes.

Absorption: about 20 percent from the stomach, 80 percent from the small intestine. Peak blood levels in 30 to 90 minutes. Food, particularly fat and protein, slows gastric emptying and substantially reduces the peak, which is the mechanism behind the advice not to drink on an empty stomach and is genuinely effective.

Metabolism, almost entirely in the liver, at a roughly fixed rate:

$$\text{Ethanol} \xrightarrow{\text{alcohol dehydrogenase}} \text{Acetaldehyde} \xrightarrow{\text{ALDH2}} \text{Acetate} \rightarrow \mathrm{CO_2} + \mathrm{H_2O}$$

Acetaldehyde is the problem molecule. It is toxic, reactive, forms DNA adducts, and is classified by IARC as a carcinogen in its own right. Your liver normally converts it onward fast enough that it does not accumulate.

The rate is fixed at roughly one unit (8 g of alcohol) per hour for an average adult, because the enzymes saturate. Nothing speeds it up: not coffee, not cold showers, not exercise, not food after the fact. This is zero-order kinetics, and it is why the only cure for intoxication is time.

Alcohol flush. Roughly 30 to 50 percent of people of East Asian descent carry an ALDH2 variant that produces a slow or non-functional enzyme. Acetaldehyde accumulates, causing facial flushing, palpitations, nausea, and headache. This is not a harmless quirk: carriers who drink despite it have substantially higher rates of oesophageal cancer, because they are exposed to far more acetaldehyde per drink. It is one of the clearest natural experiments demonstrating that acetaldehyde is the carcinogen.

In the brain, alcohol enhances GABA (the main inhibitory transmitter) and inhibits glutamate (the main excitatory one). The result is sedation, disinhibition, impaired coordination, and impaired memory formation. A blackout is a failure of the hippocampus to write memories while the person remains conscious and functional, which is why it is possible to hold a conversation and retain nothing.

2. Units and standard drinks

Different countries define these differently, which makes guidance hard to compare.

Calculating UK units: $\text{units} = \dfrac{\text{ABV %} \times \text{volume in mL}}{1000}$

UK guidance: no more than 14 units a week for both men and women, spread over three or more days, with drink-free days. That was revised downward in 2016, and the accompanying statement was that there is no level of drinking that can be described as safe. Canada went further in 2023, advising that risk begins at any level and that 3 to 6 drinks a week carries moderate risk.

3. What changed in the science

For thirty years, the standard finding was a J-shaped curve: moderate drinkers had lower all-cause and cardiovascular mortality than abstainers.

Three problems dismantled it:

1. The sick quitter effect. Studies commonly classified as "abstainers" people who had stopped drinking, often because of illness. That makes the abstainer group artificially unhealthy and the moderate drinkers look good by comparison. Studies that separate lifelong abstainers find the protective effect shrinks substantially or disappears. A 2023 JAMA Network Open meta-analysis by Zhao and colleagues, covering 107 studies, found that once these biases were corrected, no significant reduction in mortality risk was found for low-volume drinkers.

2. Mendelian randomisation. This method uses genetic variants that affect alcohol metabolism (and therefore consumption) as a natural randomisation, avoiding the confounding that plagues observational studies. Large MR studies, including work in Chinese and UK populations, find that genetically predicted higher alcohol intake is associated with linearly increasing cardiovascular risk and blood pressure, with no protective threshold.

3. Confounding by everything else. Moderate drinkers in Western cohorts are, on average, wealthier, better educated, more socially connected, and healthier than both abstainers and heavy drinkers.

The current position, reflected in the 2018 Lancet Global Burden of Disease alcohol analysis covering 195 countries: the level of consumption that minimises harm across all causes is zero, though the increase in risk at very low levels is small.

What has not changed: heavy drinking is unambiguously harmful, and that has never been disputed.

4. What alcohol actually does to the body

Cancer. IARC classifies alcohol as a Group 1 carcinogen, and it causes cancer of:

The mechanisms are acetaldehyde forming DNA adducts, oxidative stress, impaired folate metabolism, and, for breast cancer, raised circulating oestrogen.

There is no safe threshold for cancer risk. Risk rises from the first drink; it just rises slowly at first. Public awareness of the alcohol-cancer link is strikingly low: surveys repeatedly find that fewer than half of people know alcohol causes cancer, and awareness of the breast cancer link is lower still.

Liver. Progression is typically fatty liver (reversible, occurs in most heavy drinkers within weeks) → alcoholic hepatitis (potentially fatal) → cirrhosis (largely irreversible, with risk of liver failure and hepatocellular carcinoma). Women develop liver damage at lower intakes and shorter durations than men.

Cardiovascular. Raises blood pressure dose-dependently, causes atrial fibrillation ("holiday heart"), and at high intakes causes cardiomyopathy. The MR data suggest the blood pressure effect begins at low intakes.

Brain. Chronic heavy use causes shrinkage and cognitive impairment. Wernicke-Korsakoff syndrome comes from thiamine deficiency, which alcohol causes by impairing absorption and depleting stores; this is why intravenous thiamine is given before glucose to anyone with alcohol dependence presenting confused, since glucose consumes the last of the thiamine and can precipitate the disaster (Chapter 15).

Sleep. Alcohol shortens sleep latency and then wrecks the rest of the night: it suppresses REM sleep in the first half, causes rebound in the second, fragments sleep, and worsens sleep apnoea by relaxing airway muscles. The subjective experience of a nightcap helping is real; the measured sleep quality is worse.

Pregnancy. Fetal alcohol spectrum disorder is entirely preventable and permanent, causing growth restriction, facial features, and lifelong cognitive and behavioural impairment. No safe level in pregnancy has been established, and guidance in the UK, US, and elsewhere is to avoid alcohol entirely when pregnant or trying to conceive.

Nutrition. 7 kcal per gram, and effectively no nutrients. A bottle of wine is around 600 kcal. Alcohol also displaces fat oxidation, because the body prioritises clearing it, and increases appetite.

Interactions. This belongs with the medicines section and is worth flagging here:

• Paracetamol plus chronic heavy drinking is a serious combination for the liver (Chapter 65).
• NSAIDs plus alcohol raise gastrointestinal bleeding risk.
• Benzodiazepines, opioids, sedating antihistamines, gabapentinoids: additive respiratory depression and sedation. This combination kills people.
• Metronidazole and a few others cause a disulfiram-like reaction: violent nausea, vomiting, flushing.
• Metformin: raised lactic acidosis risk with heavy drinking.

5. Hangovers

Multiple mechanisms, no proven cure:

• Acetaldehyde accumulation as the clearance system saturates.
• Dehydration, because alcohol suppresses vasopressin, promoting urine output.
• Congeners, the non-ethanol compounds produced in fermentation and ageing. Darker drinks (bourbon, brandy, red wine, dark rum) contain far more than clear ones (vodka, gin), and controlled comparisons find worse hangovers from bourbon than vodka at the same alcohol dose.
• Inflammation, sleep disruption, and gastric irritation.

What has evidence: eating before and while drinking, alternating with water, choosing clear spirits, drinking less, and time. What does not: hair of the dog (it postpones), most hangover supplements, and "sweating it out."

6. Making sense of it

Some things that are true at once:

• Alcohol has real harms with no safe threshold for cancer, and public awareness of this is poor.
• The harms at low levels are small in absolute terms. One drink a day increases breast cancer risk by roughly 7 to 10 percent relative, which for most individuals is a small absolute change.
• Most drinking-related harm is concentrated in heavy drinking, and the largest single public health gain would come from reducing that.
• Alcohol has genuine social and cultural value to a great many people, and public health advice that ignores this is ignored in turn.
• The reasonable framing is informed choice: know that it is a carcinogen, know your intake, and decide.

If you want to reduce it: count units honestly for a week first (almost everyone underestimates), set drink-free days rather than a daily maximum, use smaller glasses (a 250 mL wine glass is a third of a bottle), alternate with water, and note that alcohol-free beer and wine have improved enormously and are now a genuinely viable substitute in a way they were not a decade ago.

If you drink heavily and want to stop, do not stop abruptly without advice. Alcohol withdrawal in dependence can cause seizures and delirium tremens, which has a real mortality rate. Medically supervised withdrawal exists for this reason. This is the one substance in this book where suddenly stopping can be more dangerous than continuing.

7. The bottom line

• Ethanol is metabolised to acetaldehyde, a carcinogen, at a fixed rate of about one unit an hour. Nothing speeds it up.
• The apparent heart protection from moderate drinking has largely dissolved once the sick-quitter bias was corrected and once genetic studies avoided the confounding entirely.
• Alcohol is a Group 1 carcinogen causing cancer of the mouth, throat, oesophagus, liver, colon, and breast, with no safe threshold. Breast cancer risk rises measurably from about one drink a day.
• UK guidance is under 14 units a week for everyone, with drink-free days, and an explicit statement that no level is safe.
• No safe level in pregnancy. Alcohol flush from ALDH2 deficiency is a warning sign, not a curiosity, and drinking through it raises oesophageal cancer risk substantially.
• Alcohol wrecks sleep architecture even when it makes you fall asleep faster, and it interacts dangerously with paracetamol, NSAIDs, and every sedative.
• If you are alcohol-dependent, stopping abruptly can be dangerous. Get medical support.

👉 That completes what you eat and drink. Next, the medicine cabinet, starting with what a drug actually is.