The objective of this study was to determine the population pharmacokinetic parameters and the extent of absorption of 2',3'-dideoxyinosine, a nucleoside analog with activity against human immunodeficiency virus in vitro and in vivo, after oral and intravenous administration through the use of NON-linear Mixed Effects Modeling. The data were drawn from the pharmacokinetics section of an open-label, multicenter phase I study. One center administered ddI on a once-daily schedule. The other centers administered the drug once every 12 h. Drug was administered intravenously, and the plasma concentration-time profile was determined. Patients were then given the drug orally at twice the dose used in the intravenous portion of the study, and the pharmacokinetic profile was again determined. A 40-fold range of doses was examined. Forty-six human immunodeficiency virus-infected patients were studied. Concentrations in plasma were determined by high-pressure liquid chromatography. Clearance of the drug from plasma was 47.7 liters/h/70 kg of body weight. The terminal half-life was 1.4 h. The volume of distribution in the central compartment was 18.8 liters/70 kg. Absorption was rapid, with an absorption half-life of 0.52 h. Bioavailability with once-daily administration was 27%. For twice-daily administration, bioavailability rose to 36%. This difference was significant (P much less than 0.01). For doses of less than or equal to 5.1 mg/kg given every 12 h (10.2 mg/kg/day), bioavailability was 41%. We conclude that once-daily administration results in lower mean bioavailability, probably because of a saturation of the absorption process similar to that seen with acyclovir. This difference in bioavailability on the basis of the administration schedule explains the different short-term maximal tolerated doses identified in phase I trials of this agent.