Pharmacodynamic modeling of thiopental anesthesia

Abstract

We have pharmacodynamically modeled the relationship between the thiopental serum concentration and its effects on the electroencephalogram (EEG). Power spectral analysis was used to calculate the spectral edge, a measure of the underlying EEG frequency that characterizes the progressive slowing of the EEG induced by thiopental. Eight male volunteer subjects had venous thiopental serum concentrations measured, and 10 surgical patients had arterial serum concentrations measured. Thiopental was infused at a rate of 75 to 150 mg/min until a burst suppression EEG pattern was evident. Frequent blood samples were obtained during and after the infusion for measurement of serum thiopental concentrations, and the EEG was recorded for subsequent off-line power spectral analysis to calculate the spectral edge. With venous blood sampling, it was not possible to demonstrate significant hysteresis between the thiopental serum concentration and the spectral edge, allowing thiopental concentrations to be directly related to the spectral edge. With arterial blood sampling, significant hysteresis was present, requiring an effect compartment to relate concentration to effect. The half-time for equilibration (mean ± SD) between concentration and response for the arterial data was 1.2± 0.30min. This value for K_eo is consistent with known values for cerebral blood flow and thiopental brain: blood partition coefficient. Arterialvenous concentration differences cause the apparent lack of hysteresis with venous blood sampling. An inhibitory sigmoid E _max pharmacodynamic model optimally characterized the relationship between thiopental concentrations and the spectral edge. This model allows estimation of the thiopental serum concentration that causes one-half of the maximal EEG slowing (IC₅₀), which is a measure of an individual's sensitivity to thiopental. Except for the hysteresis, there was no statistical difference in the parameters of the inhibitory sigmoid E _max pharmacodynamic model when venous and arterial blood samplings were compared. Arterial blood sampling offers some distinct advantages when pharmacodynamically modeling continuous, rapidly changing measures of drug effect, such as the EEG.