A comparison of parametric and non-parametric approaches to target-controlled infusion of propofol.

Nineteen adult patients of either gender received intravenous infusions of propofol, scaled to estimated lean body mass (LBM), for 150 minutes as part of a balanced anaesthetic. Arterial blood was assayed for whole blood propofol. The first subject received propofol at a fixed rate of 0.058 mg x min(-1) x kg(LBM)(-1). Subsequent groups received variable rate infusions based on the ratio of the infusion rate to the propofol concentration at each sampling point in the previous group, multiplied by the target concentration. After groups of one, two, five and 11 subjects, the median weighted residual was 0.040 and median absolute weighted residual was 0.153. Population pharmacokinetic analysis of the final group of six females and five males, aged 29 to 70 years and of 16.5 to 44.2% body fat, resulted in a two compartment pharmacokinetic model with coefficients and standard errors of V = 0.102 (0.0155) l/kg(LBM), V2 = 0.257 (0.079) l/kg(LBM), k10 = 0.423 (0.069)/min, k12 = 0.222 (0.051)/min, k21 = 0.084 (0.02)/min and clearance = 0.0418 (0.0023) L x min(-1) x kg(LBM)(-1). The only significant covariate was LBM. Within infusion data improved prediction when compared with data derived in previous studies from random observations.