Prediction of pharmacokinetic parameters and the assessment of their variability in bioequivalence studies by artificial neural networks

Purpose. The methodology of predicting the pharmacokinetic parameters (AUC, c(max), t(max)) and the assessment of their variability in bioequivalence studies has been developed with the use of artificial neural networks. Methods. The data sets included results of 3 distinct bioequivalence studie s of oral verapamil products, involving a total of 98 subjects and 312 drug applications. The modeling process involved building feedforward/backpropa gation neural networks. Models for pharmacokinetic parameter prediction wer e also used for the assessment of their variability and for detecting the m ost influential variables fur selected pharmacokinetic parameters. Variable s of input neurons based on logistic parameters of the bioequivalence study , clinical-biochemical parameters, and the physical examination of individu als. Results. The average absolute prediction errors of the neural networks for AUG, c(max), and t(max) prediction were: 30.54%, 39.56% and 30.74%, respect ively. A sensitivity analysis demonstrated that for verapamil the three mos t influential variables assigned to input neurons were: total protein conce ntration, aspartate aminotransferase (AST) levels, and heart-rate for AUG, AST levels, total proteins and alanine aminotransferase (ALT) levels, for c (max), and the presence of food, blood pressure, and body-frame for t(max). Conclusions. The developed methodology could supply inclusion or exclusion criteria for subjects to be included in bioequivalence studies.