Population Pharmacokinetics Of Azithromycin By NONMEM

Population pharmacokinetics (PopPK) is a rather new branch of pharmacokinetics (PK) developed in recent decades. Population pharmacokinetics combines the characters of pharmacokinetics model with population statistics concepts. The variance in the patients’ population can be described quantitatively based on ordinary pharmacokinetics’ fundamental principles and statistics method, and thus the relationship between individual’s characteristics and population pharmacokinetics can be established. The PopPK method would be very helpful in the clinical reasonable medication, and has been proposed by the U.S. Food and Drug Administration to be an essential part in the submission of the innovative drugs. But at now days in China, the studies related to PopPK have not been widely conducted yet, and related reports are rather rare.Azithromycin is used as the model drug in this study. Drug blood samples from multiple bioequivalence studies approved by the State Food and Drug Administration (SFDA) of China were set as the fundamental data base. The modeling via population approach of oral-dosed azithromycin in health Chinese male volunteers was conducted by NONMEM method so as to promote reasonable medication and control of side-effects.HPLC/MS was used to develop a sensitive and accurate analysis method in this study. Method linear range, precision, recovery and stability in different conditions were verified. The limit of detection was 2ng-mL-1, regression coefficient (r) was greater than 0.99, recovery of the azithromycin blood samples was greater than 85% and the precision (RSD) was less than 15%, respectively. The HPLC/MS method developed was proved to be acceptable for the assay of azithromycin blood samples.Because the distance between the real world and the PopPK model depends on the reasonable diversity and catholicity of the data source, a mass of azithromycin concentration data obtain from multiple bioequivalent studies were used for the establishment of azithromycin’s PopPK model and fixed factors’analysis. All of these studies were conducted under the approval of the local ethic committee and the informed consent were signed by all of the participants.The results showed that after administration of test drug and control drug, the AUC0-t、AUC0-∞、Cmax was bioequivalent in double one-side t-test. The confidence interval of Cmax, AUC0-t, and AUC0-∞were was 98.46%～113.76%,88.24%～103.24% and 88.43%～103.07%, respectively. The difference between the peak concentrations of the two preparations was not significant, means that the test drug and control drug were bioequivalent.The PopPK model was established based on the data from 8 bioequivalence studies of azithromycin oral solid formulations, totally with 160 participants. A nonlinear mixed effect model (NONMEM) method was employed to conduct the PopPK modeling of azithromycin. The final population pharmacokinetics model of azithromycin showed that the drug was absorbed and distributed in a two compartment model with first-order absorption, the clearance of the central compartment was affected by the weight factor, and the apparent volume of distribution was affected by the age factor. The inter-individual random effect distributed in an exponential way, and residual error distributed in the combination of addictive and proportional. The validation process and diagnostic plots demonstrated that the NONMEM PopPK model constructed is effective and stable.In order to evaluate the precision and stability of the PopPK model established by NONMEM method, the artificial neural network (ANN) study was conducted as well. PopPK model of azithromycin was established base on the same database. By the help of ANN Parameters Optimize System (ANNPOS), the structure parameters and the performance parameters of the neural network were optimized. After thousands times of training and adjustment of the network, the error of the ANN PopPK model was controlled under 0.001. The ANN PopPK model showed to be effective and stable via the internal and external validation.Furthermore, MSE, MAE, MDAWR, MAWR and r2 were employed to compare the precision between the NONMEM PopPK model and the ANN PopPK model.50 and 120 participants’data were randomly selected from the database and inputted into the two models simultaneously to evaluate the stability of prediction. The results showed that the MSE and MAE of ANN PopPK model is lower than those of the NONMEM PopPK model, which suggested that the ANN predicted values were closer to the detected value. As to the stability index such as MDAWR, MAWR and r2, the NONMEM PopPK model showed its advantages.These results state that the population pharmacokinetics model of health Chinese male volunteers was established successfully based on the reliable and controlled multiple bioequivalence studies. And the comparison study between NONMEM and ANN approaches was conducted, and revealed more detailed information about the population pharmacokinetics model. Due to the distinguished capability of analysis on the fixed effective and random effective factors, NONMEM approach is considered to be a more promising method to the population pharmacokinetics.