Effect of infusion rate on thiopental dose-response relationships: Assessment of a pharmacokinetic-pharmacodynamic model

Background: The rate of administration of an intravenous anesthetic induction agent is an important variable determining the total dose required to reach a given endpoint, such as loss of consciousness (LOC). The influence of infusion rate on the dose-response relationship has not been described rigorously. In this study we characterized the effect of different thiopental infusion rates on the times and doses required to reach a clinical (induction) endpoint. Methods: Fifty-six healthy, nonpremedicated men, aged 19-59 yr, were randomly assigned to receive one of seven different thiopental infusion rates (40, 60, 75, 150, 300, 600, and 1,200 mg/min). The infusion was continued until the patient dropped a held object, indicating LOC. The infusion rates were selected using a simulation which predicted the relationship between the rate of administration and cumulative dose administered at the time of LOC. Average population pharmacokinetic parameters from a three-compartment thiopental model were combined with an effect-site rate constant for thiopental equilibration of 0.58 min^-1 and a median effect-site concentration of 13.8 mg/l from previously published pharmacokinetic and pharmacodynamic models for thiopental. This derived model was used to predict the total amount of thiopental required, at each infusion rate, to produce LOC. Results: The observed median effective doses for infusion rates of 40-150 mg/min were similar and ranged from 296 to 318 mg. Dose requirements increased significantly with increasing infusion rates greater than 150 mg/min; median effective doses for infusion rates of 300, 600, and 1,200 mg/min were significantly different from each other (436, 555, and 711 mg, respectively). The original simulation underestimated the observed thiopental doses at all but the lowest infusion rate. A new simulation was performed using a recently developed combined pharmacokinetic- pharmacodynamic model. This model incorporated a four-compartment thiopental pharmacokinetic model with quantal dose-response data to derive an effect- site rate constant for thiopental equilibration of 0.29 min^-1 and a median effect-site concentration for LOC of 11.3 mg/l. The median thiopental doses predicted by this new simulation under the extreme conditions of a 30-fold range of infusion rates were within 13% of the observed doses. Conclusions: In this study we quantified the relationship between the rate of thiopental administration and the resultant cumulative thiopental dose necessary to produce LOC. This study validated a novel pharmacokinetic-pharmacodynamic model based on a four-compartment pharmacokinetic model and infusion quantal dose-response data. Finally, we demonstrated that thiopental dose-response relationships are dependent on drug administration rate, and found that the ability to predict this dependence accurately is influenced by the pharmacokinetics, pharmacodynamics, and median effect-site concentration used to simulate the dose-response relationships.