Why drinking water doesn't lower your blood alcohol
A glass of water between drinks doesn't lower your blood alcohol in any way that matters. At most it dilutes it by a percent or two for an hour or so, and it doesn't remove a single gram. There are only two ways it could work, by diluting the alcohol or by helping your body get rid of it faster, and the numbers rule out both. Water does do something real on a night out. It just happens on a different axis, and that part's worth knowing too.
Two ways water could lower your BAC
Blood alcohol is a concentration: grams of ethanol spread through the water in your body. To bring it down you'd have to do one of two things. Add enough water that the same grams sit in a noticeably bigger volume, or take the alcohol out faster than your body already does. Every version of the belief, from "it dilutes it" to "it flushes it out", is one of those two, and they fail for different reasons.
Dilution: about one percent, and only for a while
Ethanol doesn't stay in the blood. It spreads through all of your body water, which in a lean adult is roughly half to 60% of body weight (Jones 2019). That's a far bigger pool than most people picture when they imagine "watering it down".
The Watson equations, fitted to dilution measurements from 458 men and 265 women, turn that into litres from height, weight and (for men) age (Watson 1980). For a 40-year-old man of 175 cm and 75 kg:
2.447 - 0.09516 × 40 + 0.1074 × 175 + 0.3362 × 75 = 42.6 litres
A 500 ml glass of water adds 0.5 litres to that. The same grams of alcohol now sit in 43.1 litres instead of 42.6, so the concentration falls by 0.5 / 43.1, which is 1.2%. For a woman of 165 cm and 60 kg (Watson's female equation has no age term) it's -2.097 + 0.1069 × 165 + 0.2466 × 60 = 30.3 litres, and the same glass moves the concentration by 1.6%.
| Person | Total body water (Watson) | 500 ml of water as a share | Change in concentration |
|---|---|---|---|
| Man, 40, 175 cm, 75 kg | 42.6 L | 1.2% | -1.2% |
| Woman, 165 cm, 60 kg | 30.3 L | 1.6% | -1.6% |
That's the best case for water, and it isn't a real one. It assumes the glass is absorbed instantly, mixes completely and then stays. It doesn't stay. In a trial that measured exactly this, well-hydrated men who drank a litre of still water over 30 minutes passed about 1.34 kg of urine over the next four hours (Maughan 2016). The kidneys treat a water load as surplus and hand it back, and alcohol, which suppresses the hormone that tells them to hold water, doesn't make them keep it longer. The dip is temporary, and not one gram of alcohol has left your body while it lasts.
Put the 1.2% next to two other numbers. Individual blood alcohol already differs from any population model by roughly ±20-30%, so the glass moves the estimate by about a twentieth of that band (more on that band in how accurate BAC calculators really are). And the liver is removing alcohol the whole time: at 0.15 ‰ per hour, the average elimination rate the CheersCount model uses, it takes away 1.2% of a given concentration in about 4.6 minutes for every 1 ‰ of blood alcohol. The glass's entire effect on the number is a few minutes' worth of what the liver does regardless.
Water drunk at the same moment as a drink is a separate question: it can change how fast that drink is absorbed, not how much alcohol there is, and not always in the direction people expect. The post on how a Maß of beer absorbs covers why strength and volume change absorption speed.
Faster removal: the liver has a fixed speed
The second idea is that water helps you "flush it out". This one fails on how alcohol actually leaves. Between 92% and 98% of a dose is metabolised, mostly by alcohol dehydrogenase in the liver. Only 2-10% leaves unchanged through breath, urine and sweat (Jones 2019).
That enzyme is saturated early. Its Michaelis constant corresponds to a blood alcohol of only 2-10 mg/100 ml, so after the first few drinks it's working flat out, and the decline follows zero-order kinetics: a roughly constant amount per hour, whatever the concentration. In moderate drinkers about 15 mg/100 ml per hour is a good average (Jones 2010). Nothing in that depends on how much fluid passes through you. A saturated enzyme can't speed up because there's more water around it.
And the urine route is small to begin with. In experiments Jones reviews, urine collected over seven hours after neat whisky held between 0.7% and 1.55% of the dose (Jones 2019, citing his own 1990 work). Urine carries alcohol at a concentration set by the blood, so more urine does carry a little more of it out. But even if extra water doubled your urine output, which is my rough upper bound and not a measurement, the extra would be in the region of one more percent of the dose. The other 90-odd percent still goes through the liver at the liver's pace.
Someone tested it: the 2026 chaser trial
For a long time this was an argument from first principles. In 2026 it got a direct test. Imai and colleagues gave 13 healthy Japanese men sake containing 1.3 g of alcohol per kilogram of body weight over an hour, twice, in random order: once on its own, and once with 15 ml of water per kilogram sipped in between, which for a 75 kg man is over a litre. They followed breath ethanol and acetaldehyde for up to 15 hours (Imai 2026).
The curves didn't differ. Breath alcohol tracks blood alcohol closely, and the authors conclude that the water didn't significantly change how the body handled the ethanol. A litre of water, sipped right through the drinking, and the curve didn't move.
The trial was really about the hangover, though, and here the honest answer is less tidy than either side of the usual argument. Next-morning headache, nausea, concentration and sleepiness were no different with water, and neither were the scores on a timed psychomotor test, although those leaned slightly toward the water condition without reaching significance. Thirteen men is a small study. The authors say it could only reliably pick up a large effect, and it can't speak for women or for other genetic backgrounds. So: water doesn't lower the number, and whether it helps the morning after is less clear than people think. The best controlled evidence so far found no difference.
What water actually does: replacing what alcohol costs you
None of this makes water useless on a night out. Its job is fluid balance, and that's a real effect with real numbers.
Alcohol suppresses vasopressin, the hormone that tells your kidneys to keep water, so you pass more urine than the drink alone would explain. The figure usually quoted is about 10 ml of extra urine per gram of alcohol, from Eggleton's 1942 experiments. Because that loss scales with grams of alcohol while the drink itself brings fluid with it, a drink's strength decides which way it pushes you:
| Drink | Alcohol | Extra urine at 10 ml/g | Fluid in | Net fluid |
|---|---|---|---|---|
| Beer, 500 ml at 5% | 19.7 g | 197 ml | 500 ml | +303 ml |
| Wine, 150 ml at 12% | 14.2 g | 142 ml | 150 ml | +8 ml |
| Spirit, 40 ml at 40% | 12.6 g | 126 ml | 40 ml | -86 ml |
The break-even is the strength at which a drink's volume exactly pays for the urine it causes: 100 / (10 ml/g × 0.789 g/ml) = 12.7% ABV. Below that a drink brings more fluid than it costs; above it, less. Beer more than covers itself, wine about breaks even, spirits don't. A beverage hydration trial found the beer half of that directly: four hours after a litre of lager, urine output wasn't different from a litre of still water (Maughan 2016).
There's a second reason to alternate with water, and it has nothing to do with fluid. A water in place of a drink means one dose fewer, or the next one later. Spacing drinks out lowers the peak because less alcohol arrives per hour, not because the water does anything to the alcohol that's already there.
Where these numbers are weaker
The 10 ml per gram figure deserves the same scrutiny as everything above, and it comes out worse than the rest of this post.
I haven't read it off the original. Eggleton's 1942 paper exists and the citation checks out, but it's only available as scanned pages, and the 10 ml/g figure here is taken from how the paper is commonly cited rather than from the paper itself. That's secondary attribution, and I'd rather say so than dress it up.
A modern measurement comes in lower. Hobson and Maughan gave 12 men a litre of 4% beer or the same beer alcohol-free. Well hydrated, the alcoholic version produced 1279 ml of urine over four hours against 1121 ml: 158 ml more from about 31.6 g of alcohol, so roughly 5 ml per gram, half the quoted figure. When the same men started out dehydrated, the difference (261 against 174 ml) wasn't statistically significant (Hobson & Maughan 2010). The diuresis largely switches off once you're already short of water.
At 5 ml per gram the break-even moves from 12.7% to about 25% ABV, and wine becomes net hydrating too. So 10 ml/g is the cautious end of the evidence, not the best estimate, and the fluid table above is an order of magnitude rather than a measurement.
The dilution arithmetic has its own simplification: instant, complete mixing and no excretion. Every one of those makes the 1.2% look bigger than it is, so the real effect of the glass on the number is smaller still. The CheersCount model itself sizes alcohol's distribution volume with a different regression (Seidl 2000, see how the estimate works); I've used Watson here because it gives body water directly in litres.
How the model treats water
I built the CheersCount model to match the first half of this post and its water feature around the second. A water you log never touches the curve. To check, I ran a 75 kg, 175 cm man through four 500 ml beers an hour apart, with and without three 250 ml waters in between: the two curves match at every minute, peak and end time included.
Water gets its own card instead. From the second alcoholic drink of a session it shows how much water the night calls for: the larger of the fluid your drinks didn't cover at 10 ml/g and one 250 ml glass per three alcoholic drinks, rounded to 50 ml. Four beers ask for 500 ml purely for pacing, since they brought 2 litres against an estimated 789 ml of extra urine. Four 40 ml shots ask for the same 500 ml, though there the fluid shortfall alone is 345 ml. After three alcoholic drinks in a row the card adds a short "Time for a water?", and an alcohol-free beer at 0.5% or less counts as water. The card shows millilitres, never a blood alcohol number; its info link explains why water doesn't lower it. If you'd like to see the curve side, the BAC curve calculator runs the same model in your browser. It has no water field, because there's nothing for one to change.
Frequently asked questions
Does drinking water lower your blood alcohol?
No. A 500 ml glass adds roughly 1-2% to the body water alcohol is spread through, the kidneys pass it again within a few hours, and the liver removes alcohol at a fixed rate that fluid cannot raise. In a 2026 crossover trial, breath alcohol curves were the same with and without water between drinks.
Does water help you sober up faster?
No. About 92-98% of alcohol is broken down by metabolism, mainly in the liver, at a roughly constant rate per hour. Only 2-10% leaves unchanged in breath, urine and sweat, so more urine does not mean meaningfully more alcohol leaving.
Does drinking water between drinks prevent a hangover?
That is less clear than people think. In a small 2026 crossover trial in 13 men, water between drinks made no significant difference to next-morning symptoms or to a psychomotor test. The study was small and all-male, so it does not settle the question, but it does not support the popular belief either.
Why does alcohol make you pee so much?
Alcohol suppresses vasopressin, the hormone that tells the kidneys to hold on to water. The commonly cited figure is about 10 ml of extra urine per gram of alcohol, though a 2010 study measured about half that in well-hydrated men and no significant effect in dehydrated ones.
Does beer dehydrate you?
Not on fluid balance alone. A 500 ml beer at 5% brings 500 ml of fluid and costs roughly 200 ml of extra urine at 10 ml per gram of alcohol. The break-even sits at about 12.7% ABV on that figure, so wine roughly breaks even and spirits cost more fluid than they bring.
Sources
- Imai, H. et al. (2026). Water intake during drinking does not alleviate hangover symptoms. Frontiers in Pharmacology 17:1852528. doi:10.3389/fphar.2026.1852528.
- Watson, P.E., Watson, I.D. & Batt, R.D. (1980). Total body water volumes for adult males and females estimated from simple anthropometric measurements. American Journal of Clinical Nutrition 33:27-39. doi:10.1093/ajcn/33.1.27.
- Maughan, R.J. et al. (2016). A randomized trial to assess the potential of different beverages to affect hydration status: development of a beverage hydration index. American Journal of Clinical Nutrition 103:717-723. doi:10.3945/ajcn.115.114769.
- Jones, A.W. (2010). Evidence-based survey of the elimination rates of ethanol from blood with applications in forensic casework. Forensic Science International 200:1-20. doi:10.1016/j.forsciint.2010.02.021.
- Jones, A.W. (2019). Alcohol, its absorption, distribution, metabolism, and excretion in the body and pharmacokinetic calculations. WIREs Forensic Science 1:e1340. doi:10.1002/wfs2.1340.
- Eggleton, M.G. (1942). The diuretic action of alcohol in man. Journal of Physiology 101:172-191. doi:10.1113/jphysiol.1942.sp003973. The 10 ml/g figure is taken from secondary attribution, not read off the scanned original.
- Hobson, R.M. & Maughan, R.J. (2010). Hydration status and the diuretic action of a small dose of alcohol. Alcohol and Alcoholism 45:366-373. doi:10.1093/alcalc/agq029.
- Seidl, S., Jensen, U. & Alt, A. (2000). The calculation of blood ethanol concentrations in males and females. International Journal of Legal Medicine 114:71-77. doi:10.1007/s004140000154.
Every figure here is an estimate for typical people, not a measurement of you. Neither a feeling nor a calculation tells you whether you can drive.