The purpose of this study was to perform population pharmacokinetic (PPK) analysis on carbamazepine and to determine the population model of clearance of this drug in terms of individual patient characteristics. A total of 107 steady-state serum concentrations from 97 adult and pediatric epileptic patients, collected during routine clinical care, were used for the analysis. To determine the influence of different covariates on the estimate of carbamazepine clearance we used the non-linear mixed effects modeling (NONMEM) software package with ADVAN1 subroutine. This is a one-compartment model with first-order elimination and without absorption. The typical mean value for carbamazepine clearance, estimated by the base model (without covariates), in our population was 3.43 l/h. The final results of our analysis show that carbamazepine clearance increased nonlinearly with total body weight and age, and linearly with concomitant administration of valproate. The magnitude of the effect of valproate was +0.874 l/h. The interindividual variability (coefficient of variation) for clearance and the residual variability (including intraindividual variability), described by an exponential error model, were 16.76% and 31.14%, respectively. The results of this PPK analysis were validated in a group of 16 epileptic patients and suggested good predictive performance of the final model. The derived model describes carbamazepine clearance in terms of characteristics of Serbian patients, using minimal data obtained from routine clinical care of epileptic patients. This is the basis for future pharmacokinetic studies on a specific epileptic population, which will lead to better overall management of epilepsy in Serbia.

OBJECTIVES In countries with high income, tocolytic therapy with beta-mimetic agents is a cost-effective strategy compared to placebo. In our study, the cost-effectiveness of two beta-mimetic agents, ritodrine and fenoterol, used in the management of preterm labor was compared in the setting of a low-middle-income transitional country, Serbia & Montenegro. METHODS This case study was conducted at the Gynecology-Obstetrics Clinic, Clinical Center "Kragujevac," in Kragujevac, Serbia & Montenegro, between October 2004 and January 2006. In total, 235 pregnant patients with threatened preterm labor were enrolled, but 35 were lost to follow-up. Of the remaining 200 patients, 85 were given ritodrine, and 115 fenoterol. The perspective of Republic Institute for Health Insurance in Serbia was taken into account. Only direct costs were calculated; primary outcomes of the study were length of pregnancy (in weeks), time passed from the onset of uterine contractions to delivery (in weeks), and score on modified Flanagan's quality-of-life scale for chronic diseases, measured after discharge from hospital. RESULTS Prolongation of pregnancy was significantly longer in the fenoterol group (12.7 +/- 8.4 weeks) than in the ritodrine group (11.6 +/- 7.1 weeks). The mean duration of hospitalization was shorter in the fenoterol group (11.9 +/- 8.8 days) than in the ritodrine group (14.9 +/- 11.3 days). The treatment with fenoterol was less costly and more cost-effective than the treatment with ritodrine, but the difference in cost-effectiveness was not statistically significant. The cost of treatment per gained week of pregnancy prolongation was 3345.51 +/- 7668.04 CSD in the fenoterol group, and 4181.96 +/- 12,069.83 CSD in the ritodrine group. CONCLUSIONS The observed differences in treatment costs and duration of hospitalization per patient did not translate into significant differences in cost-effectiveness ratios, because of low costs of hospitalization and human labor in Serbian health system. Nevertheless, fenoterol treatment still has a tendency to be more cost-effective, and its lower acquisition cost is an advantage to this treatment option.