臨床薬理

抄録を表示する抄録を非表示にする

The population pharmacodynamics of vamicamide, a new anticholinergic drug, were analyzed by nonlinear mixed effect modeling (NONMEM) in 16 healthy male subjects. The subjects were given orally 18, 36 or 48 mg of vamicamide in a single- or multiple-dose regimen. Serum concentrations of vamicamide were measured frequently; the average urine flow rates (AFRs), estimated simply by dividing the urine volume by the time during voiding, were measured on each occasion of urination. Serum concentrations corresponding to the time of urination were predicted by curve-fitting the individual data (a total of 293 serum data) to a one-compartment model with first-order absorption. The serum concentration-AFR data (381 data pairs) thus obtained were used for NONMEM analysis. The Emax model with baseline controls was used for serum concentration-AFR response data. The voided volume significantly affected not only the baseline AFR but also maximum response (Emax). However, the concentration which causes half of Emax (EC₅₀) was independent of voided volume and was estimated to be 167 ng/ml. At the serum concentration of 140 ng/ml, which is the maximum serum concentration at therapeutic standard dosage regimen, the percent reduction of AFR from the baseline value was estimated to be 32 and 38% when the voided volume is assumed to be 100 and 400 ml, respectively. Despite the wide range of variability in AFR, the NONMEM analysis provided clinically significant concentration-response curves. This population approach appears recommendable for further clinical applications.

抄録全体を表示

PDF形式でダウンロード (1488K)

In Therapeutic Drug Monitoring, unknown plasma drug concentrations of each subject are estimated based on a few measurements obtained from the subject by a given appropriate pharmacokinetic model in which parameters have a given prior distribution. In the programs generally used for the estimation, the plasma concentration-time profile of each subject is estimated as a single curve specified by the estimated parameters of the subject, for which the modes of the posterior distribution of the parameters specified by the given measurements are often used. In the Bayesian estimation, however, the whole posterior distribution should be used to estimate the unknown plasma concentrations as a band, not a single curve, versus time.
A Bayesian approach using Monte Carlo procedures was proposed for the estimation of unknown plasma drug concentrations of a subject as a band versus time based on a single measurement obtained from the subject. In the proposed approach, an appropriate pharmacokinetic model and the prior distribution of its parameters were assumed to be given. Applications were made to real pharmacokinetic data obtained in a clinical study. The estimated bands well represented the original behavior as a whole. The results of the present study strongly suggest the clinical applicability of the proposed approach.

抄録全体を表示

J-STAGEへの登録はこちら（無料）