Two concentration-controlled trials (CCTs) defined the relationship between plasma concentrations of 3'-deoxy-3'-fluorothymidine (alovudine) and changes in surrogate markers of antiretroviral activity. In an initial open-label CCT involving 14 subjects infected with human immunodeficiency virus (HIV), unacceptable hematologic toxicity occurred when the area under the concentration-time curve during a 12-h dosing interval (AUC12) was > or = 300 ng*h/mL. Consequently, 46 subjects were assigned to AUC12s of 50, 100, or 200 ng*h/mL for up to 16 weeks in a prospective, randomized, double-blind CCT. Alovudine caused a concentration-dependent reduction in p24 antigen and peripheral blood mononuclear cell HIV titers within 4 weeks of start of treatment. The AUC12 producing a 50% reduction in p24 (108 ng*h/mL) had a trough concentration identical to the reported IC50 of alovudine in HIV-infected H9 cells. It may be possible to predict the antiretroviral activity of certain nucleoside analogues as a function of plasma drug concentration.