The effect of aging on the pharmacodynamics of midazolam was investigated in a double-blinded study involving 39 consenting patients ranging in age from 39 to 77 yr. Midazolam was infused intravenously (i.v.) using a pharmacokinetic model-driven drug infusion device to achieve a plasma midazolam concentration that was held constant for the 10-min duration of the study. Blood samples were obtained from the radial artery at 5 and 10 min for subsequent measurement of the plasma midazolam concentrations. With the 10-min sample, the patients were also assessed for the presence or absence of responsiveness to verbal command. To ensure that the pharmacodynamic end-point was assessed under the condition of a relative steady-state effect-site midazolam concentration, only those patients (n = 33) in whom the plasma midazolam concentration at 10 min was within 30% of the measured concentration at 5 min were included in the subsequent data analyses. Logistic regression was used to fit the verbal command response/no response data to a mathematical model that included patient age and the plasma midazolam concentration measured at 10 min. Cp50, the steady-state plasma midazolam concentration at which 50% of patients would be expected not to respond to a specific stimulus (e.g., verbal command), was calculated as a function of age from the parameterized logistic model. The midazolam Cp50 for response to verbal command decreased significantly (P = 0.034) with increasing patient age, demonstrating that aging increases pharmacodynamic sensitivity to the hypnotic effects of midazolam independent of pharmacokinetic factors.