Helene Peyriere(1) , Janine Bres(1) , Dominique Hillaire-Buys(2)
Laboratoire de Pharmacocinetique, Faculte de Pharmacie,
Montpeiller, France(1)
Laboratoire de Pharmacologie, Institut de Biologie, Faculte de
Medecine, Montpeiller, France(2)
Teicoplanin is a glycopeptide antibiotic related to vancomycin and is active against Gram-positive aerobic and anaerobic bacteria including methicillin-resistant staphylococci. Its structure is complex, being made up of 5 lipophilic components called A2-1, A2-2, A2-3, A2-4, A2-5, and a more polar component, called A3-1. A Bayesian approach was developed to determine the individual pharmacokinetic parameters of teicoplanin from routine clinical data. Statistical characteristics of the population parameters were evaluated using the standard two-stage method. The individual pharmacokinetic parameters of teicoplanin have been evaluated in 13 subjects by the classical maximum likelihood estimation (MLE) according to a three-compartment model. Population-typical and population-variability values were then determined for each macroconstant (coefficients and exponents) along with the variance/covariance matrix (APIS).
These population characteristics were evaluated using three pharmacokinetic softwares APIS, MICROPHARM and PKS ABBOTT (after determination of the corresponding microconstants). For some single dose data sets the outcome obtained using the MLE method (150h follow up) and that obtained using the bayesian approach with a limited sampling protocol (0.5; 1 and 24h) have been compared. The performance of the total clearance, the total volume of distribution and the terminal half-life predictions were statistically evaluated (bias, precision, and confidence intervals).
Then the bayesian estimation has allowed us to compute teicoplanin pharmacokinetic parameters from routine clinical data: on multiple dosing in 10 subjects in whom only Cmax and Cmin were available (16 samples), in 10 other subjects who had taken repeated doses of teicoplanin and with a limited sampling protocol (4 to 6 samples), in patients with renal failure under arteriovenous hemofiltration and in whom few samples were available, this only during the first day after drug administration (single dose). The performance of the teicoplanin parameters predictions with each one of these three softwares were evaluated.
The population characteristics obtained in this study can also be used for the adaptive control of the teicoplanin dosage in patients with unstable pharmacokinetic parameters due to the therapeutic (extra-corporal purification in renal failure) or to underlying pathology (obesity, anasarca).