Variability in the rate and extent of absorption, distribution and elimination of ethanol has important ramifications in clinical and legal medicine. The speed of absorption of ethanol from the gut depends on time of day, drinking pattern, dosage form, concentration of ethanol in the beverage, and particularly the fed or fasting state of the individual. During the absorption phase, a concentration gradient exists between the stomach, portal vein and the peripheral venous circulation. First-pass metabolism and bioavailability are difficult to assess because of dose-, time- and flow-dependent kinetics. Ethanol is transported by the bloodstream to all parts of the body. The rate of equilibration is governed by the ratio of blood flow to tissue mass. Arterial and venous concentrations differ as a function of time after drinking. Ethanol has low solubility in lipids and does not bind to plasma proteins, so volume of distribution is closely related to the amount of water in the body, contributing to sex- and age-related differences in disposition. The bulk of ethanol ingested (95-98%) is metabolised and the remainder is excreted in breath, urine and sweat. The rate-limiting step in oxidation is conversion of ethanol into acetaldehyde by cytosolic alcohol dehydrogenase (ADH), which has a low Michaelis-Menten constant (Km) of 0.05-0.1 g/L. Moreover, this enzyme displays polymorphism, which accounts for racial and ethnic variations in pharmacokinetics. When a moderate dose is ingested, zero-order elimination operates for a large part of the blood-concentration time course, since ADH quickly becomes saturated. Another ethanol-metabolising enzyme, cytochrome P450 2E1, has a higher Km (0.5-0.8 g/L) and is also inducible, so that the clearance of ethanol is increased in heavy drinkers. Study design influences variability in blood ethanol pharmacokinetics. Oral or intravenous administration, or fed or fasted state, might require different pharmacokinetic models. Recent work supports the need for multicompartment models to describe the disposition of ethanol instead of the traditional one-compartment model with zero-order elimination. Moreover, appropriate statistical analysis is needed to isolate between- and within-subject components of variation. Samples at low blood ethanol concentrations improve the estimation of parameters and reduce variability. Variability in ethanol pharmacokinetics stems from a combination of both genetic and environmental factors, and also from the nonlinear nature of ethanol disposition, experimental design, subject selection strategy and dose dependency. More work is needed to document variability in ethanol pharmacokinetics in real-world situations.