How Abvy estimates blood alcohol

Absorption

Alcohol from each drink is poured into a stomach compartment at a constant rate over however long you said the drink took, then absorbed into the blood at a rate that depends on your stomach's food state, from 0 (empty) to 1 (full meal).

KA_EMPTY of 4.6/hr matches the empty-stomach absorption rate Lee et al. fitted to 2,955 breath samples from 59 subjects across repeated intakes¹. KA_FULL of 1.5/hr sits within noise of the 1.45/hr fed estimate in Büsker et al.'s population model⁶.

First-pass metabolism

Not all absorbed alcohol reaches circulation. Some is broken down by stomach and liver enzymes before it ever enters the blood, a loss that grows with the food state and shrinks for people whose stomach enzyme activity is lower or absent.

The female fpFactor reflects Frezza et al., who measured gastric alcohol dehydrogenase activity and first-pass metabolism in women at 59% and 23% of men's, respectively². Bariatric bypass surgery routes alcohol around the stomach, so first-pass loss is set to zero.

Distribution

Once in the blood, alcohol distributes through your body water, not your full body weight. Abvy estimates that water volume with the Watson formula from height, weight, age and sex, then divides by the fraction of blood that is water to get a volume of distribution.

Watson, Watson and Batt derived these equations by regression from dilution-study measurements in 458 men and 265 women³.

Elimination

The liver breaks alcohol down at a rate that depends on how much is in the blood, and that rate saturates: past a point, more alcohol does not mean faster clearance, because the enzymes are already working as fast as they can.

At the concentrations most evenings reach, BAC sits well above Km, so elimination runs close to a constant rate, near the familiar 0.015 g/dL/hr figure. That is where a fixed-rate approximation still holds up. But the rate is not actually fixed: as BAC nears zero, the denominator stops shrinking as fast as the numerator, so clearance tapers off instead of hitting zero on a straight line. That is one reason "sober by" is a converging estimate rather than a simple division.

Wilkinson et al. established Michaelis-Menten kinetics for ethanol elimination from constant-rate intravenous infusions⁴. Food raises Vmax: measured at up to 50% by Ramchandani et al.'s alcohol-clamp method⁵ and at 39% (95% CI 33–46%) in Büsker et al.'s population model⁶. Abvy uses a conservative +30%.

The uncertainty band

Every parameter above is a population average, fitted to specific study cohorts, not to you. Abvy widens that average into a range by simulating the same evening at a slower and a faster personal clearance rate.

The chart's shaded band is the same evening run twice, once at each bound. It is not a statistical confidence interval from a study; it is a deliberately conservative spread layered on top of your already-personalized rate, in the same neighborhood as the roughly ±20–30% real-world variation researchers report around any single BAC estimate.

Sober time

Abvy calls an evening over once the simulated BAC, at both the slow and fast bounds, drops below 0.001 g/dL, simulating forward from your last drink until that threshold is crossed. Because elimination tapers near zero rather than cutting off, the last stretch takes longer than a straight-line estimate would suggest, which is the other reason the app shows a range of times rather than one moment.

What the app shows

Every number carries a "~" because it is a model output, not a measurement. The chart shows the uncertainty band described above instead of a single line. And "sober at" is always a window bounded by the slow and fast estimates, never a single timestamp you might mistake for a promise.

Calibration

If you log readings from a real breathalyzer, Abvy compares them with its own predictions and adjusts your personal absorption and clearance, only ever in the safer direction. Three readings across two sessions can be enough to start.

How this differs from the Widmark formula

The classic Widmark formula estimates peak BAC as dose divided by (body weight times a fixed distribution ratio r, 0.68 for men or 0.55 for women), then subtracts a constant linear elimination rate over time. It treats the whole dose as absorbed instantly. Abvy's model replaces each fixed constant with a measured relationship instead: the Watson formula for body water rather than one ratio per sex, first-order absorption spread over the time you said you were drinking rather than an instant dose, and saturating Michaelis-Menten elimination rather than a constant rate. Both describe the same physiology; Widmark trades the underlying study data for a formula that fits on a napkin, while Abvy runs the fuller model on a phone, since a phone can afford to.

Limitations

• No sensor: nothing in the app measures blood or breath alcohol directly.
• Individual variation: enzyme activity, hydration and body composition vary beyond what a handful of profile fields can capture.
• Medications and illness: liver conditions, GLP-1 medication and other conditions shift both absorption and clearance; profile flags approximate them, they do not replace medical advice.
• Tolerance: chronic heavy drinking changes clearance in ways the model does not track session to session.
• Logging accuracy: a drink that is not logged, or logged late, is a drink the model never sees.

References

1. Lee BY, Yoon HK, Baek IH, Kwon KI. Population pharmacokinetics of multiple alcohol intake in humans. Alcohol. 2013;47(2):159–165. https://doi.org/10.1016/j.alcohol.2012.12.012
2. Frezza M, di Padova C, Pozzato G, Terpin M, Baraona E, Lieber CS. High blood alcohol levels in women: the role of decreased gastric alcohol dehydrogenase activity and first-pass metabolism. N Engl J Med. 1990;322(2):95–99. https://doi.org/10.1056/NEJM199001113220205
3. Watson PE, Watson ID, Batt RD. Total body water volumes for adult males and females estimated from simple anthropometric measurements. Am J Clin Nutr. 1980;33(1):27–39. https://doi.org/10.1093/ajcn/33.1.27
4. Wilkinson PK, Sedman AJ, Sakmar E, Earhart RH, Weidler DJ, Wagner JG. Blood ethanol concentrations during and following constant-rate intravenous infusion of alcohol. Clin Pharmacol Ther. 1976;19(2):213–223. https://doi.org/10.1002/cpt1976192213
5. Ramchandani VA, Kwo PY, Li TK. Effect of food and food composition on alcohol elimination rates in healthy men and women. J Clin Pharmacol. 2001;41(12):1345–1350. https://doi.org/10.1177/00912700122012814
6. Büsker S, et al. Population pharmacokinetics as a tool to reevaluate the complex disposition of ethanol in the fed and fasted states. J Clin Pharmacol. 2023;63(6):681–694. https://doi.org/10.1002/jcph.2205
7. National Institute on Alcohol Abuse and Alcoholism. What is a standard drink? https://rethinkingdrinking.niaaa.nih.gov/how-much-too-much/whats-standard-drink
8. nidirect (UK government, Northern Ireland). Alcohol units. https://www.nidirect.gov.uk/articles/alcohol-units

Standard-drink sizes used elsewhere on this site follow the NIAAA's 14 g definition⁷ and the UK unit's 8 g definition⁸.