The pharmacokinetics of ethanol have been studied extensively since the 1930s [1, 2]. The classical Widmark one-compartment model with zero-order elimination is simple and gives a single value for the rate of elimination , but Michaelis-Menten kinetics better describe both the rectilinear part of the wash-out curve and the terminal endpoints . Therefore, models with one  or two  compartments and Michaelis-Menten elimination have been applied, although the approaches used rarely consider urinary losses of ethanol . Further development of pharmacokinetic models for ethanol have been reported in animals but the wide range of doses used cannot ethically be applied in humans [8, 9]. Difficulties in the modelling process typically consist of the variable elimination rate and the combination of linear and nonlinear kinetics. It is currently debated whether first-pass metabolism of ethanol is gastric or hepatic [10–12] and whether it is influenced by various drugs and the rate of gastric emptying [13, 14]. After oral intake, within-subject variation of ethanol kinetics is thought to be of similar  or smaller magnitude  than the between-subject variation. The components of variation, however, have not been studied after intravenous administration of ethanol.
This study was part of a larger project aimed at estimating total body water by the ethanol dilution principle. To investigate the conditions for that method, we wanted to assess the magnitude of within- and between-subject variation of the model parameters in repeat experiments by applying a two-compartment model containing parallel Michaelis-Menten metabolic elimination with first-order renal excretion to breath or blood concentration-time data and to the cumulative amount of ethanol excreted in the urine. Because analysis of breath might be preferable in many clinical settings we compared the parameter estimates obtained from invasive (venous blood) and noninvasive (expired air) sampling. The confounding factors of oral administration, such as absorption, emptying rate of the stomach and first-pass metabolism, were eliminated by using the intravenous route for administration of ethanol.