EVIDENCE LIBRARY
Does a Lower Oral Alcohol AUC Prove the Stomach Metabolized the Missing Dose?
A physiologic modeling study shows why oral-versus-IV alcohol exposure differences cannot be read directly as gastric metabolism. The rate of delivery matters when elimination is nonlinear.
No. A lower concentration–time area after oral alcohol does not directly measure how much alcohol the stomach metabolized. An original PKQuest modeling study showed that a standard oral-versus-intravenous area-under-the-curve calculation could overestimate first-pass loss when applied to ethanol. Different delivery rates and nonlinear metabolism changed the interpretation.
For DUI work, the lesson is about inference: a measured concentration curve and a calculated fraction of an administered dose are different quantities. A lower curve cannot, by itself, identify the organ responsible or justify a fixed deduction from a driver’s drinking history.
What an AUC comparison assumes
Area under the concentration–time curve, or AUC, summarizes exposure over time. In a simple system with linear clearance, comparison of AUCs after equal oral and intravenous doses can estimate systemic availability.
Ethanol complicates that calculation because its metabolic removal is nonlinear. The fraction removed during a particular interval depends on concentration and the rate at which ethanol arrives. An oral dose absorbed slowly after food is not kinetically identical to the same dose delivered by a shorter intravenous infusion.
David Levitt’s 2002 study examined that problem using a physiologically based model. It did not recruit a new drinking cohort. It reanalyzed previously published human concentration data and generated new model estimates.
What the model estimated
PKQuest represented tissues, blood flows, distribution and liver metabolism. It used intravenous data to establish parameters and then estimated the intravenous input pattern that would reproduce the observed curve after oral dosing.
The paper called this equivalent systemic input “peripheral availability.” First-pass loss was calculated as the administered oral dose minus that modeled availability. Neither quantity was directly assayed in a specimen.
The underlying datasets included breath-derived concentration information and venous measurements, with explicit model handling of concentration in blood water versus whole blood. These inputs and transformations should not be collapsed into a newly measured, uniform whole-blood dataset.
The numerical difference was substantial
For a 0.3 g/kg dose given with a meal in the reanalyzed data, the classical AUC method implied approximately 44% first-pass loss. A one-compartment approach yielded approximately 30%, whereas PKQuest estimated approximately 7%.
For the lower 0.15 g/kg dose, the main results reported approximately 73% by the AUC method, 59% by the one-compartment approach and 38% by PKQuest. The abstract gives a slightly different lower-dose model percentage, so these main-text values should be identified as such rather than presented as exact universal constants.
These are alternative estimates from particular data and assumptions. They are not statistically validated correction factors for all drinkers. Their value is to demonstrate how much the answer can depend on the model.
Why this does not locate metabolism in the stomach
Changing stomach emptying changes the speed of delivery to the intestine and liver. That can change exposure even without a large amount of gastric metabolism. The model found that the amount reaching the liver was close to the administered dose under the analyzed conditions and argued against substantial gastric removal in those examples.
That is a model-based mechanistic conclusion, not direct organ-by-organ measurement. The site’s gastric-emptying medication experiment addresses a related intervention question. Its measured curves and derived availability estimates still require separate interpretation.
Limits on applying the model to a driver
The analysis used published group means, some digitized from figures. Applying a nonlinear model to an average curve is only an approximation to an average person. Physiologic parameters were imperfectly known, and muscle blood flow was influential and adjusted during fitting.
The model also generated fed-versus-fasted simulations. Those predictions are not observations of additional participants, and the paper’s large modeled food effect should not become a blanket percentage discount for a meal.
No driving performance or individualized forensic reconstruction was validated. The findings support examining the assumptions behind an exposure calculation. They do not allow a single BAC to reveal gastric metabolism or a reported drink count to determine one exact earlier concentration.
For a separate discussion of uncertainty in calculated concentrations, see why a fixed percentage uncertainty cannot fit every calculated BAC.
Primary source
Levitt DG. PKQuest: measurement of intestinal absorption and first pass metabolism—application to human ethanol pharmacokinetics. BMC Clinical Pharmacology. 2002;2:4. https://doi.org/10.1186/1472-6904-2-4.