The population pharmacokinetics and bioavailability of oral stavudine (d4T; 2',3'-dideoxy-3'-deoxythymidine) was determined in 81 patients with AIDS or AIDS-related complex (ARC) enrolled in phase I and phase I/II dose-ranging trials. Each patient underwent inpatient pharmacokinetic studies following administration of the first oral stavudine dose; 59 patients were restudied after chronic therapy for an average of 19 days. Thirty-three of these patients also received a single intravenous stavudine dose prior to starting an oral regimen. A two-compartment model with first-order absorption and elimination was used as the structural pharmacokinetic model. A basic model provided the following population parameter estimates (interpatient variability expressed in parentheses as percent coefficient of variation): clearance/bioavailability = 30.9 (24.5%) liters/h; volume of distribution/bioavailability = 8.42 (not modeled) liters; volume of distribution at steady state/bioavailability = 68.9 (105%) liters; intercompartmental clearance/bioavailability = 12.4 (26%) liters/h; and first-order absorption rate constant = 1.32 (78.9%) liters/h. In the subset of 33 patients receiving both intravenous and oral doses, the bioavailability of stavudine was estimated to be 99.1% (18.5%). Total body weight, stage of disease (AIDS versus ARC), and an oral stavudine dose of > or = 200 mg were found to have a statistically significant but a clinically marginal effect on the estimate of the oral clearance of stavudine. This analysis shows the high degree of bioavailability of stavudine in patients with AIDS and ARC and the relatively low degree of interpatient variability in oral drug clearance compared with those of other nucleosides. Population pharmacokinetic analysis is a useful tool for assessing the combined effects of several patient variables on the pharmacokinetic properties of drugs in human immunodeficiency virus-infected patients.