Alcohol Elimination Time Calculator

Background.

Only time lowers blood alcohol. Everything else people try — coffee, a big meal, water, a cold shower, a run, fresh air, sleep — changes how alert someone feels while leaving the concentration in their blood exactly where it was. NIAAA states it directly: the body metabolises alcohol "at a steady rate, regardless of how much alcohol a person drinks or of attempts to sober up with caffeine or by other means". Feeling more awake while remaining just as impaired is the more dangerous of the two states, not the safer one, so this page has no input for any of them. It has one input that matters, and the answer is a wait.

This calculator starts from a blood alcohol concentration you already have — a breathalyser reading, a clinical or roadside result, or an estimate carried across from the BAC calculator — and works out how long it takes to fall to zero. It does not estimate a concentration from drinks; that is the BAC calculator's job, and the two pages are kept apart deliberately so that neither pretends to answer the other's question.

The answer is a range, because the elimination rate is not a constant. ANSI/ASB Best Practice Recommendation 122, the 2024 forensic standard from the American Academy of Forensic Sciences' Academy Standards Board, gives it as 0.010 to 0.025 grams per decilitre per hour, a band chosen because it "encompasses the majority of the population regardless of age, sex, ethnicity, and drinking experience", and requires that calculations "be performed using a range of elimination rates". From 0.08% — the US per-se limit, and the number most people arrive here holding — that range is a wait of three hours and twelve minutes at one end and eight hours flat at the other. Nearly five hours of difference, from one starting concentration, with no other unknowns in the model at all.

A second authority draws the band wider. Jones's 2010 evidence-based survey in Forensic Science International puts the full physiological range at 0.010 to 0.035 g/dL per hour: 0.010–0.015 on an empty stomach, 0.015–0.020 when not fasted, and 0.025–0.035 in people with alcohol use disorder undergoing detoxification, whose microsomal CYP2E1 activity is elevated. Both ranges are offered here rather than one being chosen silently. They share the same lower bound, which means the longest, most cautious wait is identical whichever you select — only the fastest-metaboliser end moves.

One limit is built into the model and reported as its own number rather than buried. Alcohol leaves the blood at a constant rate — zero-order kinetics — only while the alcohol dehydrogenase enzyme is saturated. Clause 4.1.4.3 of the standard notes that below 0.020 g/dL that may no longer hold, and clause 5.4.3 forbids forensic extrapolation below it entirely. So the last stretch down to zero is an extrapolation past the edge of the published model, and the calculator tells you how long the trustworthy part lasts.

None of this is a countdown. The times here describe how long alcohol takes to leave a bloodstream in general, not what any particular person's body will do, and certainly not the moment at which driving becomes acceptable. NHTSA's guidance is to never drive after drinking and to arrange the ride home before going out, and it notes that a person is likely to be too impaired to drive before looking — or even feeling — drunk. Impairment begins far below any legal limit, and a number on a screen is not evidence about your own nervous system.

What is alcohol elimination time calculator?

Alcohol elimination is unusual among drugs. Most substances clear with first-order kinetics, meaning a constant fraction leaves per unit time and the concentration halves at a steady interval. Ethanol does not. Alcohol dehydrogenase, the liver enzyme that does most of the work, is saturated at any concentration a drinker would notice — the Michaelis constant for class I ADH corresponds to a blood level of only 2–10 mg/100 mL — so the enzyme runs flat out and removes a constant amount per hour rather than a constant fraction. That is zero-order kinetics, and it is why a blood alcohol curve is a straight downward line rather than a decaying curve, and why doubling the starting concentration doubles the wait instead of adding one half-life.

The slope of that line is the elimination rate, written β in the literature and expressed in grams per decilitre per hour. It varies between people with liver size, lean body mass, drinking history and enzyme genetics, and within one person with time of day and recent food. Jones's survey found the slope slightly steeper in women than in men, related to differences in liver weight relative to lean body mass. None of that variation can be resolved from a form on a web page, which is precisely why the forensic standard specifies a range rather than a fitted value.

One piece of physiology sits outside this model entirely. The times here assume elimination only, which requires that all the alcohol consumed has already been absorbed. NIAAA notes that alcohol "is still absorbed into the bloodstream at a much faster rate than it is metabolized", so if the drinking finished recently the concentration may still be climbing and the clock has not started yet. ANSI/ASB 122 treats a subject as post-absorptive only when more than two hours have passed since the last drink. Start the wait from that point, not from the last sip.

How to use this calculator.

1. 01Enter the blood alcohol concentration you are starting from, and pick its unit — 0.08% is the same concentration as 0.8 g/L and 80 mg/100 mL.
2. 02Check the "starting level in g/dL" output to confirm the unit was read the way you meant it.
3. 03Choose a rate range. The forensic standard is the default; the full physiological range widens only the fast end.
4. 04Read the range, not either end alone. The slow end is the one to plan around.
5. 05Note the "to 0.020%" figure. Past that point the constant-rate model stops claiming to apply, so the remainder is extrapolation.
6. 06If your starting figure came from the BAC calculator, remember it was itself a range — these times inherit that spread as well.
7. 07Do not treat any of this as a countdown to driving. Arrange the ride before the drinking, not after.

The formula.

ANSI/ASB Best Practice Recommendation 122 §5.4.2 states the retrograde relation as AC_inc = AC_test + (β × T): to recover an earlier concentration, add back the alcohol that has been eliminated since. Rearranged for the time it takes a concentration to fall to zero, that is simply t = BAC ÷ β. There is nothing else in it. Body weight, sex, height and what was drunk all affect the concentration you start from — which is why the BAC calculator needs them — but none of them appears here, because the slope of the decline does not depend on them in this model.

The unit conversion comes first. Blood alcohol percentage is grams per decilitre, so 0.08% is already 0.08 g/dL. Grams per litre divide by ten; milligrams per 100 millilitres divide by a thousand. All three of 0.08%, 0.8 g/L and 80 mg/100 mL are the same concentration and return identical times, which the tests assert directly — a mis-selected unit is the most likely way to get a badly wrong answer here, which is why the normalised value is reported back as its own output.

Four times come out of the one division. The slowest uses 0.010 g/dL per hour, the lower bound of both published ranges; from 0.08% that is 8.00 hours. The fastest uses the upper bound of the selected range: 0.025 on the forensic standard gives 3.20 hours, and 0.035 on the Jones physiological range gives 2.29 hours. The population average uses 0.015, Jones's figure for moderate drinkers, giving 5.33 hours. Jones also reports 0.019 g/dL per hour as more appropriate for apprehended drivers "since many of these individuals are binge drinkers or alcoholics"; that figure is not used here, because a general audience is not an apprehended-driver cohort.

Because both ranges share the 0.010 lower bound, switching to the Jones range leaves the slowest figure completely unchanged and only shortens the fastest. That is deliberate and worth noticing: the conservative half of the answer — the one worth planning around — does not move no matter which authority you pick.

The fourth time is the honest boundary of the whole exercise. Clause 4.1.4.3 warns that below 0.020 g/dL the elimination rate may not be linear because zero-order kinetics may no longer apply, and clause 5.4.3 states that forensic retrograde extrapolation shall not be performed on concentrations below that value at all. So (BAC − 0.020) ÷ 0.010 is the length of the stretch the model actually claims to describe, floored at zero when the starting value is already below the threshold. From 0.08% that is 6.00 hours of modelled decline followed by a final stretch the standard declines to model.

Rounding happens once. Every division is carried at full precision, and the hour figures are rounded to two decimal places only at the point of display. The "h and min" range string is then built from those same rounded hours rather than from the raw values, so the two can never disagree — 2.29 hours always prints as 2 h 17 min, never as 2 h 17 min beside a differently derived 2.28.

A worked example.

Frequently asked questions.

References& sources.

1. [1]AAFS Academy Standards Board (2024). ANSI/ASB Best Practice Recommendation 122, 1st Ed., "Best Practice Recommendation for Performing Alcohol Calculations in Forensic Toxicology." §4.1.4.1–4.1.4.3 (zero-order kinetics, the 0.010–0.025 g/dL/h range, the 0.020 g/dL linearity limit), §5.4.2 Eq. 9, §5.4.3, §5.4.4.1, and Annex A.2 (the 0.068 g/dL case used as this page's worked example).
2. [2]Jones AW (2010). "Evidence-based survey of the elimination rates of ethanol from blood with applications in forensic casework." Forensic Science International 200(1–3):1–20. doi:10.1016/j.forsciint.2010.02.021. PMID 20304569. Source of the 10–35 mg/100 mL/h physiological range, the fasted and non-fasted sub-ranges, and the 15 mg/100 mL/h moderate-drinker average. Full text paywalled; figures cited from the published abstract.
3. [3]National Institute on Alcohol Abuse and Alcoholism. Core Resource on Alcohol — "The Basics: Defining How Much Alcohol Is Too Much." Source of "the body begins to metabolize alcohol within seconds after ingestion and proceeds at a steady rate, regardless of how much alcohol a person drinks or of attempts to sober up with caffeine or by other means", and of the note that alcohol is absorbed faster than it is metabolised.
4. [4]National Highway Traffic Safety Administration (2016). "The ABCs of BAC: A Guide to Understanding Blood Alcohol Concentration and Alcohol Impairment." DOT HS 809 844, July 2016 (Rev) — "A person will likely be too impaired to drive before looking — or maybe even feeling — 'drunk.'"
5. [5]National Center for Statistics and Analysis (2025). "Alcohol-Impaired Driving: 2023 Data." Traffic Safety Facts, Report No. DOT HS 813 713. National Highway Traffic Safety Administration — Important Safety Reminders, p.11: "The best way to prevent alcohol-impaired driving is to never drive after drinking."