How Long Until You Are Sober? Zero-Order Elimination Explained

Zero order: a constant mass per hour, not a constant fraction

Most drugs leave the body by first-order kinetics: a constant proportion of what is present disappears per unit of time, which produces the familiar exponential curve and the familiar half-life. Alcohol does not, because the enzyme that does most of the work — alcohol dehydrogenase in the liver — is saturated at almost any concentration a person is likely to reach. Once every enzyme molecule is already busy, adding more substrate cannot make the reaction go faster. The system runs flat out at its ceiling, and the ceiling is a fixed amount of ethanol per hour.

The consequence is worth stating twice because it is so counter-intuitive against everything else in a medicine cabinet. Alcohol has no half-life. There is no dose so small that it clears in minutes and no dose so large that the elimination speeds up to cope. The curve is a straight line down. Doubling the amount drunk therefore doubles the time needed to clear it: four units in a man of 80 kg take 4 h 54 to reach zero at an average rate, and eight units take 9 h 48. Compare the caffeine article in this section, where doubling the dose adds one half-life — five hours — no matter how large the starting dose was. Those are two completely different shapes of problem, and the intuitions from one do not transfer.

The Widmark relation, and the two numbers inside it

Erik Widmark set the arithmetic down in 1932, and forensic laboratories still use it. Concentration equals the dose divided by the product of body weight and a distribution factor, minus the elimination rate multiplied by elapsed time. The dose is the mass of ethanol drunk, which the standard-drink article in this section explains how to compute. The distribution factor, usually written r, accounts for the fact that alcohol distributes into body water rather than into the whole body: Widmark measured it in thirty subjects and found 0.68 ± 0.085 in twenty men and 0.55 ± 0.055 in ten women. The sex difference is a body-composition difference — more lean mass means more water means more volume for the same alcohol to spread into — and not a difference in the liver.

Work it. A man of 80 kg drinking four ten-gram units takes in 40 g, so his peak is 40 ÷ (0.68 × 80) = 0.735 g per litre. A woman of 60 kg drinking the same four units reaches 40 ÷ (0.55 × 60) = 1.212 g per litre — 65% higher from the same drinks. That is the single most useful thing the formula tells you, and it is why identical rounds produce very different states. Two caveats on the peak figure itself. It assumes the whole dose is absorbed at once with no elimination during the drinking, which over-estimates a real peak spread over an evening. And Widmark's own standard deviations on r are large: run the man's case at 0.595 and 0.765 instead of 0.68 and the peak moves between 0.840 and 0.654 g/L, which is the difference between needing 2 h 16 and 1 h 01 to fall below 0.5.

There is a unit subtlety worth flagging because it appears in the literature and confuses people. Widmark's concentration is strictly grams per kilogram of blood; blood has a density of about 1.055 kg per litre, so converting to grams per litre multiplies by 1.055. Some published forms carry the correction and some do not, which is a systematic 5.5% difference between two calculators that both look right. Every figure in this article is stated as grams per litre by the convention the market's own limit uses.

The rate, and how much it varies

Alan Wayne Jones surveyed the elimination-rate literature for Forensic Science International in 2010, and the numbers he arrived at are the ones to use. The physiological range runs from 0.010 to 0.035 grams per 100 mL per hour. Within that range, drinking on an empty stomach tends to give 0.010 to 0.015; non-fasted subjects tend to fall between 0.015 and 0.020; 0.015 remains a good population average for moderate drinkers, while 0.019 is more appropriate for apprehended drivers, many of whom are heavy drinkers; and in alcoholics during detoxification, with the microsomal enzyme system induced, the rate can reach 0.025 to 0.035.

That is a threefold spread across ordinary people, and it propagates directly into any estimate of time. Take the 60 kg woman at her peak of 1.212 g/L. At 0.10 g/L per hour she falls below 0.5 after 7 h 07. At 0.15 she needs 4 h 45. At 0.20 she needs 3 h 34. Same body, same drinks, three and a half hours of disagreement, and no way to know in advance which one applies to a given evening. This is the reason a sobriety calculator can only give a range, and the reason the range it gives should be read as the lower bound of a decision rather than as the decision.

For completeness, the worked times at the average rate of 0.15 g/L per hour. The 80 kg man on four units: below 0.5 g/L after 1 h 34, below 0.2 after 3 h 34, at zero after 4 h 54. The 60 kg woman on the same four: 4 h 45, 6 h 45, 8 h 05. The 80 kg man on eight units: 6 h 28, 8 h 28, 9 h 48 — note that the last figure is exactly twice the four-unit figure, which is the zero-order property made visible.

Nothing speeds it up, and here is why

Because the enzyme is already saturated, the rate is set by how fast the liver can regenerate its cofactor and process what it has, and nothing you do at the table changes that. Coffee is the most common belief and the most dangerous one. Caffeine antagonises adenosine receptors, which restores subjective alertness; it does not touch alcohol dehydrogenase, does not change the elimination rate, and does not remove a single gram of ethanol from the blood. The result is a person who feels considerably more capable than they are — which is worse than feeling drunk, because feeling drunk at least discourages driving.

The same logic disposes of the rest of the folklore. A cold shower raises alertness through a startle response and changes nothing chemically. Exercise and sweating remove a negligible fraction: the overwhelming majority of ethanol is metabolised in the liver, with only a small percentage leaving unchanged in breath, urine and sweat. Fresh air, vomiting after absorption is complete, and energy drinks are all in the same category. Food is the one item on the list that does something, and only if the order is right: eating before or while drinking slows gastric emptying and therefore slows absorption, which lowers the peak. Eating after the alcohol is already in the bloodstream does not accelerate its removal at all.

Six markets, six ways of writing the same idea

The table above gives each market's limit in its own units, because the units are not interchangeable and the conversions between them are not universal. American statutes define the offence in grams per 100 mL of blood and, equivalently, grams per 210 litres of breath — a 2 100:1 breath-to-blood ratio baked into the law. France, Spain and Germany work on a 1:2 000 ratio, which is why 0.5 g per litre of blood corresponds to 0.25 mg per litre of breath in all three. Portugal is the interesting exception: Article 81 of its Código da Estrada fixes the conversion at 1 mg per litre of breath equals 2.3 grams per litre of blood, so the Portuguese breath equivalent of a 0.5 g/L limit is 0.217 mg/L, not 0.25.

The reduced limits matter more than the general ones, because they catch the people most likely to be reading a page like this. Italy and Germany both go to zero for young and newly licensed drivers: in Italy under article 186-bis for anyone under 21, in their first three years of licence, or driving professionally; in Germany under §24c StVG for anyone under 21 or inside the two-year probationary period. France drops to 0.2 g/L for probationary licences and for bus and coach drivers, Portugal to 0.2 g/L for the three-year probationary regime and a long list of professional categories, Spain to 0.3 g/L for novice and professional drivers. In the United States the federal commercial limit is 0.04 and every state sets 0.02 or lower for drivers under 21. And Germany adds something the others do not: 0.3 per mille is already punishable if the driving is conspicuous or an accident follows, so the 0.5 figure is a ceiling and not a safe harbour.

Two closing cautions on the legal side. Limits change — proposals to lower general thresholds are debated regularly in several of these countries — so check the current rule in the country you are actually driving in, not the one you learned. And a legal limit is not a safety threshold: measurable impairment of reaction time, tracking and divided attention begins well below every figure in that table, which is why the reduced limits for new drivers exist at all.

Why no calculator can clear you to drive

Add up the error bars from this article. The elimination rate varies threefold across ordinary people. The distribution factor carries a standard deviation that moves a peak by more than 10% in either direction. The dose depends on pours that are routinely larger than the nominal standard drink. Absorption depends on what and when you ate. On top of that, the Widmark peak assumes instantaneous absorption, so a real curve is lower and later than the calculation. None of this uncertainty is resolvable from the inputs a web form can collect, and the errors do not cancel — they compound in whichever direction they happen to point.

So the honest use of a sobriety calculator is narrow. It is a good way to see the shape of the thing: that a heavy evening takes most of a night, that the same round lands very differently on two bodies, that the clock does not run faster because you want it to. It is not a way to decide whether to drive, and no responsible version of one pretends otherwise. If you have been drinking, the decision was made when the first drink was poured, and the answer is to get home another way.

Frequently asked questions

How long does it take to sober up after four drinks?

At an average elimination rate, a 180 lb man who has had four US standard drinks reaches zero about 6 h 43 after his peak, and falls below 0.08% after about 1 h 23. A 140 lb woman drinking the same four needs about 10 h 41 to zero and 5 h 21 to fall below 0.08%. Those are point estimates from a formula with wide error bars: the elimination rate alone varies threefold across ordinary people, so the true figure for any individual evening could be substantially longer. Treat them as illustrations, not as clearance.

Does coffee sober you up?

No. Caffeine blocks adenosine receptors, which restores subjective alertness; it has no effect on alcohol dehydrogenase and does not change the elimination rate by any amount. Your blood alcohol at any given minute is exactly what it would have been without the coffee. The dangerous part is the mismatch: a person who feels alert and is still impaired is more likely to drive than one who feels drunk. The same reasoning applies to cold showers, fresh air, exercise and energy drinks.

Why does the same round affect two people so differently?

Mostly body water. Alcohol distributes into body water rather than into the whole body, so the same grams reach a higher concentration in a smaller or leaner-mass-poorer person. Widmark's distribution factor captures this: 0.68 in men and 0.55 in women in his original measurements. Run the numbers and a 60 kg woman peaks 65% higher than an 80 kg man on identical drinks. On top of that sits the threefold variation in elimination rate, and on top of that the size of an actual pour, which is routinely larger than the nominal standard drink.

Can I still be over the limit the next morning?

Easily, and this is the situation the arithmetic is most useful for. Because elimination is zero-order, a large evening simply takes a fixed number of hours and no amount of sleep shortens it. An 80 kg man who drank eight ten-gram units needs about 6 h 28 to fall below 0.5 g/L and 9 h 48 to reach zero at an average rate — and at the slow end of the published range, considerably longer. Someone who stopped drinking at two in the morning can be over a limit at nine. If you drove home on that logic and the market you are in applies a zero or 0.2 g/L limit to new drivers, the margin is smaller still.

Why is the legal limit written differently in each country?

Because each legal system picked a unit and a breath-to-blood conversion and wrote them into law. American statutes use grams per 100 mL of blood or per 210 litres of breath, a 2 100:1 ratio. France, Spain and Germany use a 1:2 000 ratio, so 0.5 g/L of blood is 0.25 mg/L of breath in all three, with Germany writing the blood figure as 0.5 per mille. Portugal's Código da Estrada fixes 1 mg/L of breath as 2.3 g/L of blood, which makes its breath equivalent 0.217 mg/L. The numbers are close, the conventions are not identical, and a breath reading legal in one country is not automatically legal in the next.

Is there a level at which alcohol does not affect driving at all?

No health or road-safety authority claims one. Measurable effects on reaction time, tracking and divided attention appear well below every legal limit in the table above, which is exactly why so many countries impose zero or near-zero limits on new and professional drivers and why Germany can already penalise 0.3 per mille where driving is conspicuous. A legal threshold is an enforcement line drawn for practical reasons, not a point below which alcohol stops doing anything. If you intend to drive, the useful number is none.