The pharmacokinetics and pharmacodynamics of adinazolam mesylate (10 mg), N-desmethyl adinazolam mesylate (NDMAD, 10 mg), and alprazolam (1 mg) were investigated in 9 healthy male subjects in a randomized, blinded, single-dose, 4-way crossover study. All drugs were intravenously infused over 30 minutes. Plasma adinazolam, NDMAD, and alprazolam concentrations, electroencephalographic (EEG) activity in the beta (12-30 Hz) range, performance on the Digit Symbol Substitution Test (DSST), and subjective measures of mood and sedation were monitored for 12 to 24 hours. Mean pharmacokinetic parameters for adinazolam, NDMAD, and alprazolam, respectively, were as follows: volume of distribution (L), 106, 100, and 77; elimination half-life (hours), 2.9, 2.8, and 14.6; and clearance (mL/min), 444, 321, and 84. More than 80% of the total infused adinazolam dose was converted to systemically appearing NDMAD. All 3 benzodiazepine agonists significantly increased beta EEG activity, with alprazolam showing the strongest agonist activity and adinazolam showing the weakest activity. Alprazolam and NDMAD significantly decreased DSST performance, whereas adinazolam had no effect relative to placebo. Adinazolam, NDMAD, and alprazolam all produced significant observer-rated sedation. Plots of EEG effect versus plasma alprazolam concentration demonstrated counterclockwise hysteresis, consistent with an effect site delay. This was incorporated into a kinetic-dynamic model in which hypothetical effect site concentration was related to pharmacodynamic EEG effect via the sigmoid E(max) model, yielding an effect site equilibration half-life of 4.8 minutes. The exponential effect model described NDMAD pharmacokinetics and EEG pharmacodynamics. The relation of both alprazolam and NDMAD plasma concentrations to DSST performance could be described by a modified exponential model. Pharmacokinetic-dynamic modeling was not possible for adinazolam, as the data did not conform to any known concentration-effect model. Collectively, these results indicate that the benzodiazepine-like effects occurring after adinazolam administration are mediated by mainly NDMAD.