A physiologically realistic model is used to provide insight into the design of sampling protocols for accurate determination of AUC(0-infinity) for drugs subject to enterohepatic cycling. Through simulation of plasma concentration-time curves for such drugs it is found that more than one peak is predicted after oral and intravenous administration of a single dose of drug, the relative magnitude of peaks is dependent on the hepatic extraction ratio for both oral and intravenous drug administration, the percent of the AUC(0-infinity) in later time intervals is also a function of the hepatic extraction ratio, and present methods for the design of sampling protocols may not provide accurate estimates of AUC(0-infinity) (especially for highly extracted drugs), because peaks are only evident at later times after intravenous administration when plasma sampling is less frequent, much of the area occurs at later times, and the amount of drug in the sampling compartment after oral administration is much lower than that after intravenous administration of drug and could be incorrectly interpreted as low bioavailability if sampling is not carried out for a long period of time. The types of oral and intravenous profiles predicted for highly extracted drugs are exemplified by data for naltrexone in the monkey.