J-B. FAU, C. DUBRUC, S. CHAUFOUR, P. ROSENZWEIG, P. SCHEINMANN* and F. MENTRE**
Sanofi-Synthélabo Recherche, Chilly-Mazarin (F), * CIC Hopital Necker and **Hopital Bichat-Claude Bernard, Paris (F)
The pharmacokinetic profile of mizolastine, a second generation antihistamine agent, was assessed in children aged from 2 to 5 years after a single 0.10 mg/kg oral dose of the pediatric solution. Sampling times and the number of subjects with different designs were calculated from a population D-optimal protocol(1). Unfortunately only fourteen children over the 24 initially planned were included in this study (13 boys and 1 girl) with a mean (sd) age of 3.57 (1.02) years and a mean weight of 17.0 (2.68) kg. Nine subjects had 4 blood samples according to 3 different designs : at (nominal times) 0.25, 0.50, 0.75 and 6 hr or 0.25, 0.75, 2 and 6 hr or 0.5, 0.75. 2 and 6 hr. Two subjects had one sampling point at 8 hr, 1 at 12 hr, 1 at 24 hr and 1 at 32 hr. All the observed concentrations, except one, were above the LOQ. One child with a sample drawn 32 hr after drug intake had an observed concentration below LOQ which was assumed to be 0.5 ng/mL (LOQ/2). A total of 40 sampling points was available. The data were fitted with NONMEM using an open two-compartment model with a zero order absorption phase. This model included 2. A multiplicative model for the81 and 8five parameters : Tabs, V/F, CL/F, random effects was used. The error variance model was additive and multiplicative : it took into account the limit of quantitation of the analytical method (1 ng/mL). The Tabs value had to be fixed to 0.25 hr since only 3 subjects over 9 had an apparent Tabs value longer than 0.25 hr and since this was the earliest sampling point in the different designs. A first run was performed using the dose in mg/kg and estimating interindividual variability on 1 with the FO method, given previous results obtained in children 6-128CL/F and 1 was found close to zero. A8yrs old(1). The interindividual variability on 1 fixed8second run was then performed with the interindividual variability on to zero and using the FOCE method. The run was successful with an objective function of 305.6 and standard errors estimated for all population parameters. The population mean values and estimation coefficients of variation were V/F = 2 =81 = 0.713 h-1 (6.8%) and 80.274 L/kg (5.5%), CL/F = 0.139 L/h/kg (9.8%), 0.117 h-1 (21.4%). The interindividual coefficient of variation on CL/F was 2 corresponded to a coefficient ofF26.9% (33.6%). The estimated value of variation of 21.2% for high concentrations with an estimated error of 42.2%. In this run in mg/kg, no relationship with body weight or age was found in the POSTHOC estimate (CL/F). Furthermore, when a run was performed using the total dose (mg) with variability only on CL/F and using the FOCE method, the objective function was larger (319.9). The final model was therefore the one with dose in mg/kg and interindividual variability only for CL/F. The goodness of fit plots for this model were satisfactory. These parameters are close to those obtained in 6-12 yrs old children.
(1) : Population pharmacokinetic analysis and optimisation of the experimental design for mizolastine solution in children. France Mentré, Catherine Dubruc, Jean-Paul Thénot, Journal of pharmacokinetics and pharmacodynamics, in press.
Reference: PAGE 10 () Abstr 245 [www.page-meeting.org/?abstract=245]
Poster: poster