This study was performed to establish an experimental method for the investigation of interactions between ethanol and drugs under predictable and controlled conditions. The model was tested with flumazenil (Ro 15-1788), a short-acting benzodiazepine antagonist with an elimination half-life of 1 h. Six healthy volunteers (5 males, 1 female) were administered ethanol by intravenous infusion with stepwise changing rates. The infusion rates were adapted to each subject on the basis of individual disposition parameters of ethanol, which were derived from preceding short-term infusions of 120 min duration (1.0 mg/kg in males, 0.8 mg/kg in the female). This two-step procedure led to individual ethanol plasma levels between 1.47 +/- 0.04 and 1.71 +/- 0.03 g/L, which were reached after 2.5 h and thereafter maintained over another 6 h. Within the period of constant ethanol levels, single doses of flumazenil and placebo, respectively, were injected intravenously as a bolus (2 min) in a double-blind fashion according to a randomized two-way crossover design. Three subjects received a dose of 0.10 mg/kg of flumazenil, and the remaining three subjects received a dose of 0.20 mg/kg. Evaluation of the plasma concentration time curves of flumazenil did not reveal evidence of an effect of ethanol on the pharmacokinetics of this drug.