Evaluation of bioavailability by different methods.

Demonstration of bioavailability (or bioequivalence) is a condition for marketing of a drug product, unless a waiver is granted. Biovailability is usually based on some statistical evaluation of three parameters, the area under the concentration-time curve, AUC, the peak concentration, Cmax, and the peak time, tmax, obtained for both the test drug product and the standard, using a crossover design with 12 to 24 subjects. The three biovailability parameters can be obtained by various mathematical procedures, and may vary according to the procedure selected and the operator's decision. The purpose of this study was to compare two of the most widely used evaluation procedures, a curve-fitting method (NONLIN 84) and the linear trapezoidal rule, with a recently developed compartment model and operator input independent procedure, KINPAK. As the model drug, theophylline was used after single dose administration (experimental sustained-release tablet vs elixir) and multiple dose administration (experimental sustained-release tablet vs Theodur). Theophylline exhibits not only large interindividual, but also intraindividual, variations in its disposition. Accepting the postulate that biovailability is a biologic quality control test, KINPAK not only recognizes outliers of concentration-time data within a sequence of individual data sets but also outliers within a sample of individuals. Neither for the single dose study nor for the multiple dose study were significant differences found between the various evaluation methods. This cannot be guaranteed for other studies. KINPAK offers primarily the following advantages, standardized computation of all pharmacokinetic parameters derivable from a concentration-time course without making arbitrary model assumption, more or less questionable data points are recognized, omitted, but listed, automatic tabulation and/or plotting of raw data and evaluation results without transcription errors, ready for submission to regulatory agencies.